<%BANNER%>

Developing Strategies for Customized Manufactured Housing

Permanent Link: http://ufdc.ufl.edu/UFE0021914/00001

Material Information

Title: Developing Strategies for Customized Manufactured Housing
Physical Description: 1 online resource (58 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: Building Construction -- Dissertations, Academic -- UF
Genre: Building Construction thesis, M.S.B.C.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Manufactured housing is assumed to be an improvement of the construction industry compared to the traditional site-built process. Industrialization of the construction process created many benefits including greater productivity, higher quality and reduced time and cost of construction. However, many obstacles exist facing the progress of manufactured construction such as the inability to integrate more advanced technologies and automation in addition to the negative effect of customization on the production process: although customization of the process is feasible, including it in the process limits the benefits that might be achieved by industrialization, transforming the process into a labor intensive process. Simulation is one of the most effective tools to test management strategies and their effect on the production line in order to decide on the adequacy of using such strategies in the real-life process. This research will analyze one of the most widely used existing simulation tools, Stroboscope, and its ability to model customization and the dynamic decision making process in the manufactured housing industry. Guidelines of features and capacities to be incorporated into existing or new tools will be produced allowing these tools to model the stream of orders and dynamic decision strategies and outcomes as a first step towards validating and testing these strategies used in the manufactured housing construction process.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (M.S.B.C.)--University of Florida, 2008.
Local: Adviser: Flood, Ian.
Local: Co-adviser: Issa, R. Raymond.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0021914:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021914/00001

Material Information

Title: Developing Strategies for Customized Manufactured Housing
Physical Description: 1 online resource (58 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: Building Construction -- Dissertations, Academic -- UF
Genre: Building Construction thesis, M.S.B.C.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Manufactured housing is assumed to be an improvement of the construction industry compared to the traditional site-built process. Industrialization of the construction process created many benefits including greater productivity, higher quality and reduced time and cost of construction. However, many obstacles exist facing the progress of manufactured construction such as the inability to integrate more advanced technologies and automation in addition to the negative effect of customization on the production process: although customization of the process is feasible, including it in the process limits the benefits that might be achieved by industrialization, transforming the process into a labor intensive process. Simulation is one of the most effective tools to test management strategies and their effect on the production line in order to decide on the adequacy of using such strategies in the real-life process. This research will analyze one of the most widely used existing simulation tools, Stroboscope, and its ability to model customization and the dynamic decision making process in the manufactured housing industry. Guidelines of features and capacities to be incorporated into existing or new tools will be produced allowing these tools to model the stream of orders and dynamic decision strategies and outcomes as a first step towards validating and testing these strategies used in the manufactured housing construction process.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (M.S.B.C.)--University of Florida, 2008.
Local: Adviser: Flood, Ian.
Local: Co-adviser: Issa, R. Raymond.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0021914:00001


This item has the following downloads:


Full Text





DEVELOPING STRATEGIES FOR CUSTOMIZED MANUFACTURED HOUSING


By

HAOURAA DIALA DANDACH















A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE IN BUILDING CONSTRUCTION

UNIVERSITY OF FLORIDA

2008

































2008 Haouraa Diala Dandach


































To my father, Rakan Dandach, and my mother, Sabat Daher, as an act of love and appreciation









TABLE OF CONTENTS



L IS T O F F IG U R E S ....................................................................................... ............................6

A B ST R A C T .............................................................. ............................. ................ .... 7

CHAPTER

1 IN T R O D U C T IO N ......................... ............................................................................. 9

Overview of the Construction Industry ................................... .... ...................... 9
Advantages and Disadvantages of Manufactured Housing ...................................... 10
A im O objective and Scope .................................................................... ..................... 11

2 LITERATURE REVIEW ..................................................................................................... 13

Manufacturing Industries and Similarities with Manufactured Housing .......................... 13
Site-Built Construction and Differences with Manufactured Housing................................ 14
M manufacturing Supply Chain................................................................ ..................... 14
L ean C construction .................................................................................. ....... .......... 17
Improving the Efficiency of the Production Process in Manufactured Housing............... 18
Improving the Efficiency the Material Flow and Management System in Manufactured
H housing ...................... .. ... ...... .... .. .. .. .. .... ........ ....... ...... .......... 19
Modeling and Simulation of the Manufactured Construction Process................................ 20

3 M E T H O D O L O G Y ............................................................................... ............................23

Existing Mapping of the Production Process.................................................. 24
Customization of the Manufactured Construction Process .............................................. 26
The Management and Decision Making in a Customized Construction Process................ 27

4 INCORPORATING CUSTOMIZATION AND MANAGEMENT STRATEGIES
INTO THE PRODUCTION PROCESS........................................................29

Customizing the Production Process ............................................................ 30
Managing the Manufactured Construction Process ...................................................... 33
O v erv iew ...................................................................................... ............... .......... 3 3
D defining the D decision ................................................................... .............................. 34
Causes, Strategies and Outcomes of Decisions ............................................................. 35

5 CONSTRUCTION SIMULATION TOOLS AND FLEXIBILITY IN MODELING
DYNAM IC DECISION M AKING ....................................................................................42

Stroboscope's General Characteristics for Dynamic Modeling .......................................... 42
Strategies for Decision 1: "Do We Start Incoming Orders or Should We Wait?"......... 44
Stroboscope's General Shortcomings in Simulating the Strategies........................... 44









Analysis and Modeling of Reduced Process ............................................................... 45
Reduced Model: Simulation Language Deficiencies .............................. ........... 46
Capacities of Adequate Simulation Tools ............................................................. 47

6 C O N C L U S IO N S .................................................................................. ............................50

R research R results .................................................................................................................... 50
Conclusions and Recom m endations .................................................... ........................... 52

APPENDIX: REDUCED PROCESS MODEL: STROBOSCOPE'S LANGUAGE...................54

L IST O F R E FE R E N C E S ...............................................................................................................55

BIO G RA PH ICA L SK ETCH ........................................ ..................................................... 58









LIST OF FIGURES


Figure page

3-1 M ethodology Flow Chart Representation.......................................................................24

4-1 Custom ized Production Process........................................................ ..............................31

4-2 EzStrobe Model: process map of customized orders, Order Type A ..............................32

4-3 EzStrobe Model: process map of customized orders, Order Type B...............................32

5-1 Simplified Customized Process ......................... ...............................................45

5-2 Flowchart for Requirements of Decision 1 Implementation...........................................48









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science in Building Construction

DEVELOPING STRATEGIES FOR CUSTOMIZED MANUFACTURED HOUSING

By

Diala Haouraa Dandach

May 2008

Chair: Ian Flood
Cochair: R. Raymond Issa
Major: Building Construction

Manufactured housing is assumed to be an improvement of the construction industry

compared to the traditional site-built process. Industrialization of the construction process

created many benefits including greater productivity, higher quality and reduced time and cost of

construction. However, many obstacles exist facing the progress of manufactured construction

such as the inability to integrate more advanced technologies and automation in addition to the

negative effect of customization on the production process: although customization of the

process is feasible, including it in the process limits the benefits that might be achieved by

industrialization, transforming the process into a labor intensive process. Simulation is one of the

most effective tools to test management strategies and their effect on the production line in order

to decide on the adequacy of using such strategies in the real-life process. This research will

analyze one of the most widely used existing simulation tools, Stroboscope, and its ability to

model customization and the dynamic decision making process in the manufactured housing

industry. Guidelines of features and capacities to be incorporated into existing or new tools will

be produced allowing these tools to model the stream of orders and dynamic decision strategies









and outcomes as a first step towards validating and testing these strategies used in the

manufactured housing construction process.









CHAPTER 1
INTRODUCTION

Industrialization of the construction process was introduced in the early part of the

20th century through the construction of manufactured housing. This innovation in the

construction industry had many benefits over the standard on-site construction including

greater productivity, higher quality and reduced cost. In 1998, around 22% of new single-

family housing was constituted of manufactured housing. This new industry managed to

provide the American home buyers with an affordable alternative offering quality, safety,

cost-effectiveness and duration reduction.

Overview of the Construction Industry

In the past years, housing demand had risen considerably. In 1998, the housing

demand was roughly short of 2 million housing units/year (Willenbrock 1998). Today,

American cities are still suffering from shortage in affordable housing. Eighty per cent of

the 1,000 large and small American cities surveyed by the National League for Cities in

2007 reported that rising housing costs are putting a severe strain on families. For

example, Chicago (Population: 2.9m) identified an immediate need of at least 200,000

affordable units: in Minneapolis, Minnesota (Population: 383.000) over 50,000 units

were identified and in Lodi, California (Population 67,000) identified a need of 8,000

units (City Mayors Society 2008). In face of the growing housing crisis which mainly is

an affordable housing crisis in the past two decades, the construction industry is faced

with many challenges: the traditional site-built has been reluctant to use innovation and

technology to improve efficiency and maintain quality. This is due to its unique supply

chain, site variability and the risks extending beyond the contractual liability into the risk

of reputation. Manufactured housing (MH) on the other hand is becoming more appealing









to the American consumer especially since manufactured homes, unlike site-built homes,

can benefit from mass purchasing of materials, products and appliances, adding savings

to the cost of purchase of the homebuyer.

Advantages and Disadvantages of Manufactured Housing

Manufactured housing is considered as one of the cheapest solutions at a rate of

$40.80 per SF compared to a $91.99 per SF for site-built single houses in the year 2006

(MHI 2008). Approximately 65 corporations constitute the MH industry with 230

factories throughout the United States (MHI 2008). In addition to it being a cost-effective

housing solution, MH housing has other advantages: for instance, a higher level of

quality can be achieved since the construction is undertaken in a controlled environment

including labor supervision, control of all aspects of the process and weather-free

interference which prevents from delays. Better production rates and reduced

construction time and cost are the results (MHI 2008).

Despite this fact, the number of manufactured homes was noted to decline from

approximately 350,000 units in 1999 to 117,000 units in 2006 (MHI 2008). This decline

in industry can be reported to the fact that the MH industry is still primitive in its

technology and modem equipment use compared to other manufacturing industries

(Barriga 2003). Several constraints exist limiting the technology breakthrough of

manufactured housing and mostly the fact that it is labor intensive and does not really

incorporate many repetitions especially in the case of customization (Senghore et al.

2004). As a consequence, innovation and application of new devices to improve both the

production process and material engineering areas become vital to the MH industry

(AbuHammad et al. 2004).









In an attempt to improve the efficiency of the process, different aspects have been

researched and analyzed such as the production process, the material flow and its

management. Approaches such as lean production and supply chain have been studied

and adopted to generic manufacturing systems covering the process from the suppliers to

the customers. Attempts to provide better productivity were discussed to cover parts of

the process and simulation was used to represent the processes mainly by network

representation. Another main reason for the manufacturing industry decline would be the

lack of customization: although possible and successfully adopted in the industrialized

process, the customization posed a different kind of difficulties. The more customization

and design options are available, the less the industrialized process becomes efficient

resulting in more overheads, coordination problems and idle time.

However, throughout the literature review, very few papers were found

mentioning the customization of manufacturing process and the development of generic

models for decision making that will include customized orders.

Aim, Objective and Scope

This paper's overall aim is to define and include customization in the generic

representation of manufactured construction and represent the decision making process in

the modeling of such a process.

The objectives of this research are represented as follows:

Identify the decisions being made and their components and define a generic
decision making process representative of a manufactured housing construction
process allowing for customization in the modules being built.

Review of the existing construction simulation tools and their adaptability to
represent the dynamic decision making process: Stroboscope, an advanced
construction simulation language, will be analyzed for its efficiency and
flexibility in representing such decisions and their outcomes.









Suggestions concerning requirements from new construction simulation tools to
better represent customization and the decision making process.

The approach in this research is qualitative as to describing the decision making

process and will be limited to finding of one site visit to a manufactured housing plant in

Lake City, Florida. The construction process was found to include very few aspects of

industrialization due to the management strategy used and to the few incoming orders:

this created a labor intensive process and the productivity was not as sensitive to decision

strategies as expected. The main strategy was first in first out. In addition to the limitation

of the case study findings, the research is also limited to only part of the construction

process and does not include the overall supply chain.









CHAPTER 2
LITERATURE REVIEW

Manufacturing Industries and Similarities with Manufactured Housing

In an attempt to study the possible approaches to better represent and analyze the

customized construction process, a review of existing approaches that have been applied

to similar processes in other industries was performed. Complex systems were

approached in a variety of methods depending on the complexity, precision required and

mainly the level of human interference. In cases where only qualitative approaches

desired, fuzzy logic has been used. This approach is related to the fact that complexity

and imprecision are two incompatible factors especially once the human intervention and

analysis are involved: for instance, a fuzzy controller can be designed to replace a human

operator in control systems. Complex systems are also approached with fuzzy logic like

the evaluation of the performance of gas turbines (Center and Verma 1998). This method

however describes the variables in qualitative rather than quantitative terms, which

although applicable in areas like economics, bioengineering and traffic, can have several

downfalls when used in the construction industry (Center and Verma 1998).

A closer resemblance exists between the car manufacture industry and

construction, mainly the manufactured housing construction where labor move from one

product to the other in a fragmented process. Standardization, modularization have been

used to simplify the process and increase productivity in an environment where there is

no variation in material flow, labor, production and demand. More importantly,

customization seems to be the main barrier for the enhancement of the production process

in the project-oriented site built homes and the main reason of underachievement of

manufactured housing (Wiesel 2004).









Site-Built Construction and Differences with Manufactured Housing

The site-built construction is the most popular construction method to date.

However, this does not lead to conclude that it is the most efficient and productive

construction method: in fact, it is in fact one of the most labor intensive and least

productive processes; quality and productivity are difficult to control due to the excessive

variability of the environment, the human factor (labor, management, subcontractors,

suppliers...) and many other variables that affect the overall process. In fact, even when

implementing the same project at close by locations using the same contractors, the cost,

schedule and quality are never similar. In order to improve the site-built process, some

suggest a more modular framework for the structural and service systems included in the

construction process, separating the systems. This is expected to allow for more

innovation and the introduction of technology in the actual construction process with the

use of manufactured components. The goal would be a product with more

interchangeable parts, lower cost, better quality, faster, higher quality and lower

maintenance costs (Bashford 2004). It remains that more similarities exist between the

manufactured housing process and other manufactured industries than with the site-built

construction process.

Manufacturing Supply Chain

Supply chain management (SCM) is the process of planning, implementing, and

controlling the operations of the supply chain as efficiently as possible. Supply Chain

Management extends over all movement and storage of raw materials, inventory, and

finished goods from the origin to the consumption. Some researchers hypothesize that, in

general, a key factor to successful business is through managing the entire supply chain.

Some researchers in the manufactured housing (MH) industry started focusing on









studying the efficiency of the entire supply chain in comparison to the traditional

manufacturing and quality based improvements.

Coordinated planning of inventory-distribution integrated systems is the earliest

version of SCM and started as early as in the 60s. Research gradually progressed to

include first two or more stages of the supply chain and finally to models incorporating

the three traditional stages in the supply chain: procurement, production and distribution.

Decoupling departments or functions (assembly, storage and distribution) of the same

facility has proved to be an inefficient and non-competitive decision making policy

(Thomas and Griffin 1996).

A study of 215 North American manufacturing firms revealed the causal linkages

between sourcing decisions, manufacturing goals, customer responsiveness and

manufacturing performance revealed the following: strategic sourcing decisions (strategic

outsourcing and supplier management) influences the degree of manufacturing goal

achievement (dependability, flexibility, cost and quality) which in its turn influences the

level of customer responsiveness. However, the statistical tests show no strong

correlation between the manufacturing goal achievements positively influencing the

degree of manufacturing performance. This can be explained by the fact that the labor

productivity index used should be replaced by another index of manufacturing

performance such as investment productivity (Narasimhan and Jayaram 1998).

Consequently, simulation was researched as a tool for analysis and evaluation of

supply chain design and management alternatives. Smith, Sadeh and Swaminathan (1998)

base their framework on three main manufacturing industries: car manufacture, computer

manufacturer and grocery industry. Similar processes were identified despite the









differences and two main categories were defined: structural elements modeling

production and transportation of products and control elements specifying various control

policies. Supply chain interactions require more dynamic and sophisticated controls than

the first in, first out (FIFO) queues. However, their framework and model basis did not

include continuous manufacturing simulation or decision-making controls modified with

the evolving conditions (Swaminathan et al. 1998).

Strategic alliances of independent (and international) companies created the

Integrated Supply chain Network (ISN). Modeling is a basis for effectiveness, efficiency

and optimizing cost, time and quality: strategic and capacity planning models are

nonlinear integer programming models whereas operational level decision making

models use discrete event models (Viswanadham 2000). With the evolution of the SCM

beyond the multi facility unique enterprise boundary into the cross-enterprise scope,

technology was pushed a step further to allow for adequate cross enterprise simulation

that preserves confidential internal information of the concerned companies (Gan et al.

2000).

Supply Chain in Manufactured Housing: For years, SCM for MH industry focused

only on the manufacturing and quality aspects of the process. Jeong, Hastak and Syal

(2006) were the first to approach the SCM for MH as an entire supply chain using 5

manufacturers and 5 retailers in Indiana: the principle parties were defined to be the

customers, the retailers, the manufacturers and the suppliers. For each group, the various

characteristics and flows were detailed, analyzed and then grouped: the material and

information flow were combined showing all relations between the parties (Jeong et al.

2006).The conceptual supply chain model set the framework and basis for developing a









simulation model in order to measure the performance of supply chain systems in place

or possible improvements of new alternatives (Hastak et al. 2006). The model is divided

into four sub-models: house ordering, site and foundation preparation, house production

in factory and house install and set-up. A dynamic, discrete and stochastic simulation

model ARENA was used and the following assumptions were made: double-wide house

order, exponential probability distribution for customer generation and triangular

distribution for the other processes, and the material availability is consistent. The

performance measures selected were (1) how many house orders are generated in one

year, (2) how long it takes a house from order to installation, and (3) the average queue

waiting time for each process. The main bottleneck was found to be from the time

difference between the site foundation preparation time and the house manufacturing

time: this leads to the conclusion that processing time at the factory needs to be

improved. The scenario of a web-based ordering system was analyzed and found to

optimize the SCM system (Hastak et al. 2006).

Lean Construction

By definition, lean construction is concerned with the "holistic pursuit of

concurrent and continuous improvements in all dimensions of the built and natural

environment: design, construction, activation, maintenance, salvaging, and recycling". In

construction, this approach attempts to improve construction processes by maximizing

value and minimizing cost and accounting for customers' needs. This accomplished

through minimizing waste of materials, time, and effort in order to generate the

maximum possible amount of value (Koskela et al. 2002). Designing a production system

to achieve the stated ends is only possible through the collaboration of all project









participants (Owner, A/E, Constructors, Facility Managers, End-user) at early stages of

the project, going beyong the traditional forms of contracts used in today's industry.

One of the applications of this principle to MH is the lean assembly where the

process is simplified by industrialization, modularizations, standardization, and

continuous flow processes, yielding to reduced waste and higher quality.

The lean TFV (Transformation, Flow and Value) theory applied to construction

management in the manufactured context defines it as the allocation of resources

transforming inputs into outputs while maximizing flow and value to the customer

(Abdelhamid 2004). In a conventional construction management model, the critical

performance measure is the capacity utilization. In a lean construction model, planning

reliability is thought to increase system production. In fact, Chitla and Abdelhamid

(2003) compared the conventional labor utilization factor (LUF) and the percent plan

complete (PPC): the first one revealed problems like lost time due to measurements or

travel whereas the latter reveled that the main issue was in the workflow between stations

in addition to the individual production station. Improvements suggested by the first

included adding jigs and more inventory which will only create additional cost whereas

the latter suggested improvements included labor issues such as work pace, skill and

education, material supply, unclear directions and other issues which, when addresses

correctly, will improve the throughput of the system (Chitla and Abdelhamid 2003).

Improving the Efficiency of the Production Process in Manufactured Housing

Senghore (2004) analyzed the production process in the MH industry and defined

it with five major areas: floors, walls, roofing, exterior finishes and interior finishes. The

process was however represented partly and does not include all the activities and

stations from materials to finished product. In contradiction with the traditional









construction management practices, it was found that improving the utilization of the

independent stations does not improve the overall production (Senghore et al. 2004).

However, optimization of utilization of resources is still a common approach for

many manufacturing companies and for researchers. To attain more flexibility and better

resource utilization, the facility layout design was studied in its present most common U-

shape and was found to include much inefficiency (Mehrotra et al. 2005). Banerjee

(2006) evaluated the layout design based on the material flow in a quantitative approach

based on optimizing material travel distances and costs (Banerjee et al. 2006). Some of

the strategies for streamlining the production process concluded by Abu Hammad (2004)

were to minimize the number of stations and create substations and split and merge

activities so that the number of labor required will be consistent to finish work at each

station in addition to automating the movement from one station to another (Abu

Hammad et al. 2004).

Improving the Efficiency the Material Flow and Management System in
Manufactured Housing

The material flow and management system was not researched or studied with

equal interest or attention until recently. This was due to the subjective scenarios in each

factory transforming the production line into an efficient material management process.

According to (Barriga et al. 2005), the existing material flow and management systems

consist of a material flow system and an inventory control system. The inventory control

system allows for customized materials and in addition uses an independent demand

system in opposition to the dependant system suggesting lean production and just-in-time

inventory control systems. In any case, the generic material flow and management system

was developed by defining and incorporating all parties (suppliers, dealers and









customers) and departments (purchasing, sales and production) involved. This generic

system allowed for material requirement estimation and to relate demand for materials to

the master production schedule noting however the drawbacks mentioned previously

(Barriga et al. 2005).

In addition, using a data base approach, an efficient material requirement planning

system was developed managing the all data involved in material requirements

estimation, allowing for better planning strategies and flexible with the changing demand

(Jeong et al. 2005).

Modeling and Simulation of the Manufactured Construction Process

Static and floating bottlenecks along the production lines were found to be the

main cause of delays along with other negative implications such as frustration and

exhaustion since workers pace have to continually vary (increasingly or decreasingly)

eventually affecting the quality of the product. This has affected the will of manufacturers

to be innovative with new material and custom designs, affecting tremendously their

competitive edge. The impact of bottlenecks is better evaluated by simulation: however,

the variable management process restricts the accurate simulation (Mullens 2004).

The construction industry has not yet adopted simulation in its decision making

process mainly due to the complexity of the process and to the extensive time required to

correctly simulate a process. Most applications still remain in the academic research

level. General purpose simulation, usually presenting the user with more flexibility

allowing for accurate representation, requires users to be knowledgeable in simulation

language. Some researchers argue that special purpose simulation addressing a specific

sector of the industry will provide a user-friendly tool that will be adopted (AbouRizk

and Hajjar 1998).









Stochastic modeling is the common representation of variation. However this

approach does not deal with the causes of the variation and could cause inaccuracy of the

simulation results. The Artificial Neural Networks (ANN) approach is a tool capable of

learning complex relationships between input and output data. The output of the ANN

can be used as parameters to the simulation model. Such a generic integration model

allows the external system to exchange information with the model for instance allowing

factors like soils condition, soil classification, excavator and truck relative positioning to

affect production in an excavation simulation model (Hajjar et al. 1998).

In discrete event simulation, it is assumed that the state of a system changes at

specific times marked by certain event. Discrete event simulation is very adequate to

model construction processes. These simulations can be performed using general purpose

simulation and special purpose simulation. Although general purpose simulations are

more flexible since they can be manipulated to better suit the requirements of the process,

they require a lot of work and simulation knowledge which makes them less appealing in

the industry. Special simulators and simulation languages are more domain-specific and

have several approaches: Event scheduling (ES), activity scanning (AS) and process

interaction (PI).

Several simulation systems use diagram based on activities or activity cycle

diagrams (ACD) represented by networks consisting of nodes (activities and queues)

connected with arcs or links. These systems clock advance and apply both the AS

strategy combined with ES for event generation mechanisms. Many examples of

simulation tools or programs exist such as: Stroboscope, Cyclone cyclic operation

network, Hocus hand or computer universal simulator and many others (Martinez 1996).









Stroboscope is a very advanced simulation language/tool capable of modeling

very complex systems. However, it remains very difficult to use since it requires

knowledge of the language used. EzStrobe (Martinez 2001) was developed to facilitate

the use of simulation tools with a user-friendly modeling graphical interface. It uses the

basic Cyclone ACDs, Stroboscope simulation software and the MS-Visio graphical

interface. It also allows to some of Stroboscope features such as the use of variables to

parameterize the construction process and allows for standard and customized results.

However, it falls short of Stroboscope's flexibility and ability to program very complex

interactions especially the features relating to resources and their types: resources in

EzStrobe are not differentiable when residing in the same queue. However, it is still

useful for the purpose of applied research.









CHAPTER 3
METHODOLOGY

The approach to this study is qualitative in defining the problem and suggesting

alternative solutions. Initially, the existing production process will be described in detail

with all its components. In a second step, customization will be added to the existing

simulation models in attempt to represent the flexibility in options that the industry

offers. This will be done using EzStrobe, a visual and simplified version of Stroboscope

used in a previous research to represent the production process (Senghore et al. 2004).

Consequently and in a more detailed approach of the customization, specifically the

decision making process used in most MH management techniques, a detailed analysis of

the decision making process, components and strategies will be performed

simultaneously with a comparison to existing simulation tools and languages and their

adaptability to model such decisions and strategies. Given that it is one of the most

widely available construction simulation tools, Stroboscope will be mainly analyzed for

this purpose. Guidelines for developing existing tools or creating more flexible tools that

will better represent the dynamic construction decision making process will be presented

as the result of this research. The above mentioned methodology sequence is summarized

and represented in the following Figure 3-1.















LITERATURE REVIEW


CUSTOMIZATION EZ-STROBE



EXISTING
MAPPING OF
PRODUCTION
PROCESS



( ANALYSIS OF
/ DECISI ON NEED FOR MORE
MAKING ADAPTED AND
PROCESS AND
CUSTOMIZATION SIMULATION
REQUIREMENTS TOOLS








CSUGGESTION
STROBOSCOPE STROBOSCOPE








DYNAMIC DYNAMIC





DECISION
PSTROBOSCOPESS










FOR NEW
MODELING
TOOLS AND
THEIR
FEATURES










Figure 3-1. Methodology flow chart representation


Existing Mapping of the Production Process


As stated in Chapter 2 describing the literature review for this study, research


studies analyzing the MH production process exist: these studies reduced the process to a


model used for comparing and analyzing factors such as labor and material utilization in
MODELING

FEATURES









an attempt to improve the process. The process was mapped through case studies and was

assumed generic to most manufactured housing processes. In general, the process was

represented similarly to the physical findings at the MH plants: by types of works (or

activities) performed at one or more physical stations. The main five types of works were

(1) floors, (2) walls, (3) roofing, (4) exterior finishes, and (5) interior finishes. A plant

can have many stations (up to 16 stations depending on available space in the plant, on

the complexity of the homes and on the planning of activities in the various stations. In

parallel to the main production line, various feeder stations support the various activities

with resources like cabinets, fixtures and roofing supports (Senghore et al. 2004). Each

station was mapped separately showing the logic of production flow, the activities and

processes used including durations and logical sequencing, and required resources

associated with the activities including availability and quantities used (labor and

material) in addition to the stations' usage. This mapping was used to represent the

process through simulation in order to test the efficiency of the stations, resources and

determine the causes of delaying the process if any in order to improve productivity.

Limitations: However, this approach is limited by various factors resulting in an

inaccurate representation of the actual process. On the one hand, the research studies only

cover bits and pieces of the manufactured housing process: studies were limited to some

areas of the constructions (wall frames to exterior finishes) and did not include the

remainder of the process till the finalizing of the house. In addition, the areas included in

the model would not be entirely represented and broken into the detailed activities but it

was limited to a more generic approach. This is mainly due to the complexity of

approaching the subject as a whole. However, this resulted in a restricted analysis of the









system that does not truly represent the real life system. On the other hand, the

production process was approached in the representations but not the material and

resource flow and restrictions: in the actual MH plant, resources including labor crews

and materials are managed in a more flexible manner and are conditioned by the decision

of the plant production manager; in addition, the process does not account for the

diversity in customized housing and the various types of models that requires different

amounts of resources and occupies stations and activities over different durations. As a

result, since the representation of the process is not dynamic enough to represent the

flexibility and variability of the real process, the analysis and optimization of the current

modeled process cannot be considered best representative of the optimization of the real

system. In fact, the majority of the tools used in the literature for the simulation of MH

are not flexible enough to include the dynamics of the process, the customization and the

conditional decision making.

Customization of the Manufactured Construction Process

It was found that the customization once incorporated in the construction process in

MH has created many inefficiencies in the process and the productivity of the system: the

more the customization in design and construction, the less the industrialized process

becomes efficient resulting in more overheads, coordination problems and idle time.

However, the existing research did recognize that customization in the MH industry is a

requirement vital for its competing end. However, it has not been included in the models

as a major player in the optimization and output of the process: incoming orders were

considered as a continuous flow of similar models that will not affect the sequencing of

processing, timing of starting activity processing, resource usage or activity durations. In

reality, orders are not processed on a first come first serve basis regardless of their types,









numbers and the types and numbers of orders already being processed in the production

line. Decision making remains a subjective process that relies mainly on the MH plant

managers, their personal experiences and judgment. However, the process being as

complex and dynamic as it is, the need for introduction information technology

specifically computer simulation software that are tailored for this purpose has become

critical especially sine the MH industry remains inefficient when compared to other

manufactured industries such as car industries that have managed indeed to include

customization in the optimization of their processes.

The Management and Decision Making in a Customized Construction Process

For the purpose of this research, a case study of a manufactured plant in Lake City,

Florida was considered. This plant produces a diversity of models and is very flexible for

customization beyond its multitude of model plans. The process used in the plant was

considered only as basis for this research and analysis for the following reasons:

The overly customized process has reduced many of the industrial
manufacturing benefits. In fact, over-customization has left little common
components being manufactured that the process resembled the site-built
except for the controlled environment: the process was found to be very
labor intensive.
Orders have not been as abundant as expected especially with the housing
crisis taking place. As a result, large delays between orders have resulted in
a reduction in labor and space resources: many crews were laid off and only
one of the many plants in functioning. This also relieved the resources and
stations from the expected constraints in the production line and the
productivity or output (mainly the number of days per houses) was mainly
affected by the frequency of the incoming orders and not the efficiency of
the resource allocation and process management or the productivity and
usage of the resources at the stations.

As discussed, the case study will only be used as a basis for qualifying the possible

decisions to be made by the plant manager in the case of various customized orders

incoming at different time intervals. These decisions will be described in detail as far as









all the factors affecting them, the strategy and conditions that decide on the decision.

Stroboscope (Martinez 1996) is described to be the most flexible and adapted

construction simulation tool, providing its users with a dynamic simulation language

instead of a restricted template tool. This language's improvements over other simulation

tools mainly lie in the fact that it can dynamically and actively consider the state of the

simulation. This language will be discussed and analyzed to see its adequacy to represent

and incorporate the decision making process into the production process of any MH

process. Future recommendations will be made based on the deficiencies of the existing

construction simulation tools and how they could be overcome.









CHAPTER 4
INCORPORATING CUSTOMIZATION AND MANAGEMENT STRATEGIES INTO
THE PRODUCTION PROCESS

The manufactured housing production process has been extensively discussed and

analyzed as shown in Chapter 2 that describes the literature research. The process has

been described in its entirety starting from the suppliers to the homeowners and was

modeled in partitions due to the complexity of representation. The customization of the

manufactured housing process although recognized as vital for attracting homeowners

has not yet been fully integrated into process mapping or any simulation model

encountered in the literature research. Another aspect that was totally ignored was the

plant manager strategies and dynamic decisions: the orders, in addition to being

considered of only one type, were also processes as soon as they arrived. The only

restrictions on the production line were the availability of the resources, the available

openings in a station without any consideration to any strategy as far as starting the

process (lining up similar orders for instance) or managing the production line (moving

around resources or stopping one station's activities in order to allow other orders to be

processed).

As part of the objectives of this research, we will first approach the customization

of the incoming orders and include it in the model of the production process. This will be

performed using EzStrobe, a graphical simplified simulation tool that uses Stroboscope's

simulation language and Visio as a graphical interface. This simulation tool had been

successfully used to model part of the production process by Senghore (2004).

In a second more detailed approach, a series of management decisions of the

production line were defined and detailed showing all the factors affecting such

decisions, conditions for evaluations and outcomes of the strategies used. These decisions









were an estimation of possible decisions that might affect the production of the plant and

were based on observation from site visits and the literature. After defining the

components of the strategies, Stroboscope, being the most advanced construction

simulation language, was analyzed to test its flexibility to model the decision making

dynamic process: flexible aspects as well as its shortcomings were described in order to

draw conclusions as to what would the required capacities of new construction simulation

tools be.

Customizing the Production Process

In a first attempt to incorporate customization into the process and the non-

predictable incoming order variability, and based on findings during a visit to the

previously mentioned manufactured housing plant, a reduced process map was defined to

include several activities including (1) flooring framing, (2) wall framing, (3) electrical

and plumbing, (4) insulation and (5) drywall. The available stations were considered as a

resource in addition to labor (construction crews, carpenter crews, electricians and

plumbers' crews, insulation crews) and material for carpentry. EzStrobe (Martinez 2001)

is a simulation tool developed as a user friendly extension of the simulation language

Stroboscope which conserves some of its features including the following:

The use of Cyclone's basic modeling features including queues, combi
activities and normal activities;

The use of stroboscope as the simulation software while the user can
graphically use a template using MS-Visio graphical interface; and

The use of parameters and the generation of customized results.

In a first approach to the customization issue, EzStrobe was used to model and

simulate the above described process map as shown on Figure 4-1.











C-SURUC ON FL PLUMBING I N CONSTRUCTION ELEC/PLUMBIN INSULATION CONSTRUCTION
CEW CREW CREW CR CREW





FLOOR WALL INSUL N R RWALL











The main activities (all of them combis) are cutting of the floor framing, cutting

of the wall framing, construction of the floor framing, construction of the wall framing,

Electrical and plumbing works, insulation works and installation in drywalls. The

following constitutes part of the logic and conditional sequencing of the process:

The first two activities require only start with the incoming orders but are
considered parallel to the production line since they do not use the station
resources.

Floor framing is conditioned only by the activity for cutting the floor frames
and can start independently of cutting of the wall framing which conditions
the wall framing is not ready.

A continuous material generator is added in order to prevent any shortage
from material to affect the process; this, however, is not representative of
the real life scenario where at times material falls short and does create
some delay in the process.

Carpenters and materials are shared by the cutting of frames activities, with
a priority given to the cutting of the floor framing.

All other activities require labor crews and the preceding activities to be
executed.

A station space will be occupied when floor framing starts and will only be
available after the drywall is installed.












*Two types of orders are included in this model sharing the same labor,
material and space resources as shown on Figure 4-2 and Figure 4-3;
however, they included different amounts of usage of the resources and
different activity durations. In addition, the two order types are generated
through probability links shown on Figure 4-2.


Figure 4-2. EzStrobe model: process map of customized orders, order type A





i -----





I gEndFrmgb1 EdllFrmb

.FIF.gb G l rm2





Cut Ib1 Insu- tonbl 1

CuitFrFrm b1
ORDER -


Figure 4-3. EzStrobe model: process map of customized orders, order type B









The model performed the task allocated in representing the customization and

sharing resources: with the roughly assumed values, an output of around 2.5 days per

house was established. However, the use of EzStrobe was overruled because it did not

provide the flexibility required conditioning the sequence and the logic behind the

customization and it was clear that it will not be resourceful for integrating dynamic

decision making and characterizing resources.

As a result, although the interface was very easily handled, it did not provide any

of the features that are assumed available with Stroboscope and that part of the research

was only included to rule out the use of EzStrobe. The research was then diverted to

defining the decision making process and all the factors relating and affecting it in

addition to the conditional strategies and their outputs. Consequent to defining them, the

use of Stroboscope will be qualitatively assessed to check its suitability to represent the

resources, the strategies and the overall decision making process.

Managing the Manufactured Construction Process

Overview

All manufactured housing plants have a production manager or a plant manager

that actually makes decisions concerning the production and material management.

Material management was analyzed in previous research: a material supply management

system was defined that was used in improving the material requirement estimation

process (Barriga 2003). Activities were only considered in the context where they

influence the material flow and management system. The system uses backward

scheduling and introduces inventory control and supply chain management concepts to

the MH industry. However, this study did not include the economic implications through

the efficiency of the production system which still needs to be tested. In fact, processing









orders, especially when they are customized orders, on a delivery time basis only is

probably most cost efficient as far as the production throughput and resources usage are

concerned.

During the site visit to the plant, it was noted that no specific decision making

process was in place and that was mainly due to the low rate of ordering of houses. In

fact, one of the noted strategies was to randomly alternate types of ordered homes in

order to create diversity for the labor force which is supposed to be an incentive for

productivity. However, this strategy was not tested to prove its efficiency. Regardless the

reasons why such strategies are not in place, whether it is the decreasing amount of orders

due to the construction crisis or the lack of testing and control tools that allow such

decision to be virtually tested prior to their application, this remains a new field that have

not yet been researched in the MH industry.

Defining the Decision

Since no information was found in the literature or from the plant visit, several

scenarios of decisions were determined using the best judgment and common sense. After

defining a series of decisions, the factors affecting those decisions were mainly defined

from 3 types of variables: the variability of type and number of the incoming orders, the

present state of the process activities (expected ending time of activities) and resources

(usage and availability).

The following decisions were defined as guidelines for possible logical strategies

to be tested for their efficiency as far as the throughput of the production line (number of

days per house for instance):

1. Starting incoming orders or holding them.









2. Stopping processed orders and moving them out of production line or
letting them be processed.

3. Moving resources around from one station to another.

4. Produce inventory and store it.

The above defined decisions are independent headlines that cover various aspect

of the plant management. They require strategies to be determined and implemented once

the conditions for implementation are established. The outcome of the strategies needs to

be also defined as to what aspects of the process will be affected. Each decision will be

considered and analyzed independently from the others. Once these strategies are

defined, they will have to be tested at all times during the process and this will require, in

the case of a simulation, the need to access many characteristic of the simulation at all

times. And this precisely is the main obstacle faced in the common simulation tools.

Causes, Strategies and Outcomes of Decisions

Each of the enumerated decisions is analyzed separately in the following sub

paragraphs. The first decision will be analyzed and a strategy will be concluded for its

implementation and the modeling of such a strategy using Stroboscope will be discussed

in a subsequent chapter. The second decision will also be thoroughly analyzed; however,

due to its complexity and the new constraints it introduces, a strategy will not be

concluded. The other decisions will be moderately detailed as to the factors relating to

their implementation without going into further details.

Decision 1

The first decision can be phrased in a question form: Do we start incoming orders

or should we wait? In other terms, when should we start processing the incoming orders?









The factors or variables affecting the decision depend on the distribution of the

incoming orders and the current state of the orders being processed. For the incoming

orders, some of the data required are the number of orders, their time of arrival and

finishing deadlines if any. For the orders being processed at the time of the incoming

order, the required data are how many orders are being processed, what are their types

and their expected finishing time. However, the plant is not considered one block where

once the order is started it will be processed to the end. On the contrary, the production

line is composed of a series of station where a variety of independent activities are being

performed. These activities require their own resources and only share the physical space

or station with other activities. In order to have access to the expected finishing times of

any order, a manager has to have access to information such as the state of each activity:

the percentage of performed work and the remaining duration in case the activity started

and expected duration and availability of resources in case the activity is ready to start. A

previous study of the production process showed that, in contradiction to traditional mass

production paradigm, decreasing labor utilization (by adding resources to minimize

activity duration) leads to a better over all production (Senghore et al. 2004). This may

lead to assume that creating activity sets that require similar processing durations,

regardless of the number of resources needed to create that situation will probably result

in a better overall productivity of the process. Consequently, access to the remaining

duration for activity completion is necessary for the decision making process in this case.

This may also lead us to consider processing similar order consecutively for a better

efficiency of the production line: comparing the state of the production line and the

incoming orders, if the incoming order is of the same type, we might process it as soon as









an available station (in addition to material and resources) are available; if it is of a

different type, we might choose to not process that order and maybe wait to have a few

similar orders before allowing them to occupy the production line. Waiting time

restriction will be required not to exceed a certain limit if delay of a certain home will

create additional cost that exceeds any saving in the improvement of the productivity of

the process.

After determining the factors that affect this decision and the required data from the

system, the following strategy was developed in order to be eventually tested:

If incoming orders are similar to the ones being processed, then we process
the order as soon as we can.

If incoming orders are different then ones being processed and ordering
history or prediction shows there will be other similar types to that order,
then we hold the order with a different type and we process it with the
similar ones.

Do not hold orders depending on their types for more than a specified
period of time (that can be tested).

An example simplifying this strategy would be considering the production process

described on Figure 4-1 and considering two types of homes to be processed (type A and

type B): if the activities on the production line are processing all type A models and a

type B model is incoming, then we hold that type and only process new type A models.

That only happens until a point in simulation time where the type B is delayed beyond an

acceptable time and it requires being processed. At that point, we start processing all type

B models held and incoming up to the point where type A models A held are delayed

beyond the acceptable deadlines.









Decision 2

The second decision can be phrased in a question form: Do we stop orders being

processed and remove them from the production line? Or in other terms, when should we

stop orders being processed in order to allow incoming orders to start being processed?

Similarly to the first decision, the factors or variables affecting this decision depend

on the incoming order type, number and distribution in addition to the state of the

processed orders (type, numbers and activities' remaining durations). With the same logic

used with the first decision, if one processed order (or more) is creating a delay in the

production line by taking longer time for performing the activities at the different stations

which holds up the stations downstream and creates idle time for the station upstream,

then a decision must be made about whether to take that order out of the production line,

add resources to push it through the production line or just leave it on the production line.

This decision should be compared to the alternative of replacing this order by another

incoming order of similar type to the other orders being processed (which means with

similar activity processing durations). However, the decision is to stop the activities for

this order and remove it from the production line, the scenario tends to be more complex

than in the case of the first decision. The first complexity lies in the fact that access to the

state of an activity although easy but the destruction of such an activity due to a decision

process is not that easy to implement: in the most recent simulation model, once started,

an activity will be processed till its end before any condition is checked. This creates a

first obstacle in modeling such a decision with the most recent and common construction

simulation tools. If this was overcome, the second obstacle lies in how to release the

resources held by that activity before its termination in addition to the third obstacle

concerning how to reintroduce this order back into the production line and under what









conditions. The data required by this type of decisions is far more complex to obtain than

that of the first decision since it involves termination of activities being executed, release

of resources before activities are performed and introducing partly executed activities. As

a simplifying example, we can consider the production line described in Figure 4-1 and

the same conditions described for the decision 1 (type A and B models). Assuming the

two type B, one type A and two type B models being processed in that order and

considering that model A takes a longer time to be processed, we can consider removing

the type A model from the production line and replace it with an incoming type B

models.

Decision 3

The third decision can be phrased in a question form: in the case where some

activities are delaying the production line, should we move resources from one station to

the other? Or in other words, do we allow flexibility in moving resources in order to push

activities delaying the production line and keep the flow of production? This decision

concerns resources but it is conditioned by the status of the production line and the

activities being performed.

When stations start delaying the production line because they are taking longer than

the other stations, a good strategy might be to accelerate the activities in that particular

station. This can be accomplished by adding to the labor resource which will accelerate

the performance of the activities. As discussed in the second decision, with all the

impracticality of implementing it into a simulation model to be tested, the other strategy

would be to take that model out of the production line. Similarly to decision 2, the

strategy of moving resources around is not very practical to implement into a simulation

model in order to be tested. In most construction simulation tools, the only condition on









resources is their availability and required amount in order to allow activities to start.

Once the activity starts it holds the resources and will not release them until the end. In

addition, the simulation tools do not allow activities to accept resources during the

performance of the activity or in other term before the activity is completely executed.

Conditional allocation of resources in the case of a slow activity seems impossible with

the existing construction simulation tools and further research to approach this matter

should be directed toward innovation in simulating such a strategy.

Decision 4

The fourth decision is not precisely a decision since it represents a performance

strategy that will be either adopted in a plant or not and does not require a dynamic

decision making process and changing conditions. However, we will still include this in

this section since it will require similar strategies, inputs and outputs regardless of the

fact that it is not a dynamic aspect that changes along the production line.

This performance question concerns the choice of keeping inventory: will keeping

inventory increase the production flow and the overall throughput of the system by

reducing any delay that might result from lack of material inventory? Is this increase in

productivity considered profitable when comparing the increase in production (if any) to

the cost of keeping such an inventory (cost of providing space, resources required to

move material, tying capital in inventories that will wait a certain period of time before

being processed).

This strategy does not seem hard to model and test with the traditional construction

simulation tools. However, if we consider inventory to comprise not only raw material

but also modular components that could be produced for the various types of homes, the

problem starts to be complicated: preparing such modular components will require a









labor force that will not be always required to perform on the inventory activities. Such

labor crews can be considered an additional backup for the existing crews. In this case,

limitations and conditions on when to stop production modular components and move the

labor to help the other crews on the production line falls into the complexity encountered

in modeling the strategy of the third decision (moving resources around when certain

stations are delaying the production line).

In the following chapter, only the strategy of the first decision will be addressed

and its main modeling components and difficulties will be identified. As mentioned in a

previous chapter, Stroboscope (Martinez 1996) is supposed to provide its users with an

innovative dynamic simulation language. For the purpose of this research, the flexibility

of this language to model those difficulties will be discussed in a qualitative manner

providing feedback whether such a language is adequate to solve this issue or other

resources will be needed.









CHAPTER 5
CONSTRUCTION SIMULATION TOOLS AND FLEXIBILITY IN MODELING
DYNAMIC DECISION MAKING

The following chapter will consider the first decision described in Chapter 4 and

analyze the possibility of modeling it along with the customized production process using

Stroboscope. As mentioned previously, out of the widely available simulation tools for

modeling construction processes, Stroboscope is the most used and most flexible as

thought by its authors. It is thought to be the most flexible in modeling the dynamic

features of the manufactured construction processes. Stroboscope's main feature that will

be used is its characterized resources: this feature will be used to represent some of the

complex conditional resource release taking place during the decision making process.

On the other hand, other aspects of the decision making process that cannot be solved

using the features of Stroboscope will be described and possible other modeling solutions

will be suggested.

Stroboscope's General Characteristics for Dynamic Modeling

The customization process described in chapter 4 and represented using EzStrobe

in Figure 4-2 and Figure 4-3 does not efficiently represent the customized production

process. This is mainly due to the representations of the stations as resources: the stations

are not all similar and the ordering of the stations, which is not represented in the

EzStrobe model, is very important. In fact, having two similar activities sequence sharing

the same resource is not the best representation of how the production line processes two

types of orders: in the previously represented model, if enough resources are available,

the station for floor framing for instance could be occupied by two models or more before

the station resource moves to the next activity; this is not possible in the real life scenario.

In addition, this representation forces the sequence for all activities and does not allow









for flexibility of independent management of the stations: once an order is started, it will

be processed till the last activity which does not reflect the flexibility of the real scenario

where decisions at each station can be made independently.

Stroboscope's most important feature relating to our subject matter is the

simulation of unique "characterized resources" in opposition to "generic resources"

which are interchangeable and indistinguishable. Characterized resources represent more

accurately the real life uniquely identifiable resources. In addition to fixed properties that

identify and characterize certain resources, characterized resources used in Stroboscope

have other properties called "SaveProps" that change as the simulation is running and are

defined by the user according to their need. Other properties of the characterized

resources are system maintained and are very useful for identifying resources:

BirthTime which is the value of the simulation clock when at the time of creation
of resource

ResNum which is a serial number unique among resources of the same type

Timeln which contrarily to the other two is not fixed; it represents the value of the
simulation time when the resource entered its previous node; it is updated every
time a resource enters an activity or a queue.

Other properties for resources exist such as VarProp which create the ability of

creating properties that are functions of other resource properties. Stroboscope also

allows for access to system maintained statistical variables of resource properties in both

queues and activities. These properties are only available when nodes (activities and

resources) are cursored: in the case of resources in activities, access to properties is only

possible when activities are either starting or ending. In any case, in the following, we

will discuss the flexibility of stroboscope to model the decision making process in a

customized manufacturing scenario.









Strategies for Decision 1: "Do We Start Incoming Orders or Should We Wait?"

The strategy for this decision was to process similar types of orders in a sequence

in an attempt to improve productivity. If an incoming order is similar to the ones being

processed, then this order will be processed as soon as possible when the production line

allows it. If the incoming order is different than the ones being processed, this order is

held to allow for other orders to be processed. This however is limited by two variables:

the first restraint is a time limitation where an order should not be held for longer than a

certain period of time (d days). The second constraint is a limitation on the number of

orders allowed to be held: after a certain number (n) of held orders of a certain type, those

orders can then be processed in a sequence.

Stroboscope's General Shortcomings in Simulating the Strategies

In order to achieve this logic, access to certain data is required:

Orders should be characterized and differentiated as to their type, the time
and date of receipt and their waiting time before being processed.

The orders being processed at the various stations (and activities): access to
the type of the order resource that is activating a certain activity (combi).

Stroboscope allows for several attributes for characterized resources that allow for

a determined and flexible resource acquisition for instantiating activities through drawing

resources from queues; these attributes include: ordering of the resources in a queue

according to a specified "discipline", an order drawing through the draw links and sorting

when drawing (DrawOrder), drawing under certain conditions from a queue

(DrawWhere) and drawing until a certain condition will be met either in the queue or in

the instantiated combi (DrawUntil). One of the benefits of DrawOrder is that it can

regulate drawing of resources for two different activities requiring different sets of

conditions. However, such a sorting affects greatly the simulation time.









Similarly to its drawing feature, release features of the characterized resources have

the same behavior: the same type of conditional release applies to the release of resources

from the activities that are terminating using attributes such as ReleaseOrder,

ReleaseWhere and ReleaseUntil.

Analysis and Modeling of Reduced Process

In an attempt to test Stroboscope's flexibility to model the first decision scenario,

the problem was broken down into its simplest aspect in an attempt to simplify it for the

purpose of this study. Along the same logic describing the production process in the

previous chapters, resources represented by two different types of orders are considered

along with only two activities. The resources are initiated as five resources of type A and

two resources of type B. Order resources are qualified by their type and processing

durations. The process is represented on Figure 5-1.




ORDERS ACTIVATE ---PROCESS END



Figure 5-1. Simplified customized process

The different conditions for drawing orders were defined according to the types and

count of the resources. The conditions are detailed as follows:

As a first general rule, when the order type in "Activate" and "Process" are
similar, then draw the order with the same type from the queue Orders.

If the orders in "Activate" and "Process" are not similar, then check the
number of orders of each type in the queue: Start processing the one with
the largest number.

Never keep an order for longer than 2 days without being processed.









The first obstacle was in the fact that duration of activities is not flexible depending

on the resource type: The duration pf processing one type is different than the duration of

processing another; hence duration of the activity and the type of the order should be

connected.

Reduced Model: Simulation Language Deficiencies

For the purpose of this paper, the main features of Stroboscope were studied in an

attempt to capture its main abilities to represent and simulate the customization in the

MH process and the decision making process. It is to be noted that due to many

restriction, this study acknowledges its limitations as to the full understanding of the

details of the simulation tool at hand: this is due to the time constraint along with the

complexity of the language itself that requires deep and lengthy evaluation. In

acknowledging this fact, we highlight the possibility of overlooking and failing to use

some possible features that could have been used to work around obstacles found in

representing the researched process with Stroboscope.

The simplified customized process represented in Figure 5-1 was represented using

stroboscope. The difficulties were encountered in modeling the logic of the strategies

enumerated in the preceding paragraph where the use of the features was not as flexible

as expected and lead to errors and misrepresentations.

The first difficulty encountered was in the assumption that we can at least order the

resources by their waiting times. However, this is not possible since using the

"Discipline" function, it will evaluate a resource only one at the time it enters the queue

and will not reevaluate to check if this resource has been waiting for 2 days (as required).

In order to compensate, the function "filter" was used; the filter forms a subset of

resources within the queue but it still did not solve the issue.









Another related matter to this one is in the generation of various distributions of the

same resource of different types: although, using "Generate" allows for generation of

resources during the time when the simulation is running, this creation of resources is on

the one hand restricted by the start and end of events (such as start or termination of an

activity) and on the other hand such a generation cannot be randomized in the case of

Stroboscope's characterized resources (it could be in the case of generic resources).

When trying to compare the type of resources contained in the activities, it was

impossible to do so in Stroboscope: using "DrawWhere" which allows selecting

resources that satisfy certain conditions; this is due to the inability to compare resources

that are in activities in the case where the resources are not cursored. In fact, the use of

this function, in such a case, could lead to a run time error which interrupted the

simulation.

The software is attached in Appendix A to show the attempts made to run the

simulation. However, the overall impression was the inability of Stroboscope to simulate

the complex and dynamic decision making process even in a simplified process like the

previous one. Restraints are due to the aspects of the decision making process that need

access to certain information about the resources such as the types of models being

processed and what type of these resources are being processed.

Capacities of Adequate Simulation Tools

Based on the previous discussion regarding the deficiencies and limitations of the

existing construction simulation tools, and based on the requirements of the

manufacturing plant for decision making strategies and implementation, a basic flowchart

for the conditions, processes and required input data for the implementation of Decision 1









was represented as shown on Figure 5-2. The incoming orders are basically checked for

three main conditional decisions prioritized in the following sequence:

The first condition refers to resource delays in the queue holding the orders:
if orders have been delayed for more than a specified allowed value, then
these delayed orders will be processed. If all delays are still within limits of
time delay, then orders can still be held in queue and other conditions for
processing will be checked.

The second condition refers to activities in the production line and the
orders that they are processing: if they are processing orders similar to any
of the incoming orders, then this type of order will be processed; if not, the
third condition will be checked.

In case none of the previous conditions were fulfilled, a third condition for
processing should be checked: independently of time, orders held and not
processed should not exceed a certain specified number. When incoming
orders held of a certain type exceed this number, orders of that type should
be processed by the order of their incoming time.


Figure 5-2. Flowchart for requirements of decision 1 implementation









These strategies for decision require access of specific data concerning activities

and resources at all times and not just during the time where an order is incoming or an

activity is being processed (activated). Such a language or tool should give access to

statistical data cross-referenced with the simulation time.

The above mentioned logic is intended as a basic guideline for requirements of a

new simulation tool or language or adjustments to be performed on existing simulation

tools such as Stroboscope. The strategies will require an adequate simulating tool that

will effectively represent the conditions and make the required data available. Once

properly modeled, these strategies should be tested as to their efficiency in improving the

production. Other strategies will also be represented, simulated and compared and the

best strategies will be used for real life implementation to improve plant productivity.









CHAPTER 6
CONCLUSIONS

This study provided a basic analysis of the current state of the manufactured

construction industry in referenced to customization. After a brief analysis of the benefits

of such customization and based on the literature, a set of decisions and their respective

strategies were defined to incorporate customization and to be simulated and tested. In a

parallel manner, a more detailed analysis on existing simulation tools was performed to

analyze their flexibility in simulating such strategies. As a result, the deficiencies of the

existing construction simulation tools were established and recommendations for

improving those tools to better represent the construction process were suggested in an

attempt to ameliorate and upgrade the simulation process and eventually the efficiency of

the testing strategies in order to implement them in real-life. This chapter will include the

results of this study and the conclusions and recommendations provided for future

research.

Research Results

Following the sequence defined in Chapter 3 describing the methodology and steps

followed, various results were established at each step of this study. First, during the

preliminary research of the construction manufacturing industry, the importance of

customization was established as the most defining element for the competitive edge of

the industry in comparison with the rather unique sit-built construction. However,

allowing for too much customization, although easily applied in manufactured

construction, will produce many obstacles to the industrialization process itself, creating

a labor intensive process that no longer benefits from automation of the process or other

benefits of industrialization. On the other hand, and throughout the literature reviewed









during the course of this research, it was found that the production process in

manufactured housing was analyzed and modeled independently of the customization: in

fact, most of the studied reviewed did not include or allow any customization in the

process and only one type of homes was being processed. Consequently, all the aspects of

management and decision making were not incorporated in those studies. This represents

on of the major holdups that the manufactured housing industry has been suffering from,

compared to other industries such as car manufacturing where automation and simulation

have been largely introduced and used to improve the processes.

These results lead to the second part of this study where a set of dynamic decisions

and strategies were defined and detailed to represent possible scenarios that need to be

introduced in modeling and simulating the production process in a manufactured

construction plant. Concurrently, an analysis of the existing tools and their flexibility to

model such dynamic decisions was performed: the case study of one of the most widely

available construction simulation tool, Stroboscope, was used. Several of its

characteristics were found to be flexible and representative of many aspects of the

construction industry such as the "characterized resources" which allow for uniquely

identifiable resources with properties that vary with the simulation time and represent the

state of the resources. However, many deficiencies in this simulation tool/language also

exist when it comes to representing the dynamic customized manufactured construction

decision making process. The difficulties encountered during the modeling of the basic

process proposed in Chapter 5 are representative of the barriers that the language contains

in representing the dynamic decision making process. Such difficulties are the

impossibility of ordering the resources in a queue after the resources have been









introduced into the queue, the incapacity to define releasing strategies that use the

statistics of resources in queues and finally the language inability to access resources'

statistics in activities unless the activities are cursored.

Conclusions and Recommendations

The findings and results of this study concluded that the available construction

simulation tools were not able to represent the dynamic decision making process existing

in the real life customized construction process. Amelioration of the existing construction

simulation tools or new approaches to the dynamic decision making modeling are

required to benefit the industry and add to its competitiveness especially with the hit that

the housing industry has recently taken and the need for more affordable and good quality

housing.

This research sets some basic guidelines for the logic that a construction simulation

language would be expected to model in order to benefit from simulating the processes

and gaining directives for better more productive management strategies. These

guidelines remain shy of representing a global solution for the obstacles facing the

construction manufacturing industry. Simulation in many other manufacturing industries

has shown its efficiency even in a customized environment such as car manufacturing:

allowing for customization did not put at risk the process productivity and did not affect

negatively the cost, time or quality of the products. Another main manufacturing industry

having more comparable aspects to the construction industry is the computer

manufacturing: various components are integrated in all products in addition to product

specific components such as motherboards that vary from one product to another. This

industry relies tremendously on automation and uses various aspects of the automation









gains. Future research in manufactured housing can analyze the simulation concepts used

to solve the complications of modeling this dynamic process.

It remains that the construction industry in general is a very complex process that

includes many variances and variables and extends upstream and downstream the

production line to include the Client (customers) and the Suppliers. Such variances work

against the complete automation and fine tuning of the manufacturing system. In

addition, customization in the construction of homes creates many variations on different

levels that add to the disturbance of the industrialization of the construction process. As a

conclusion, manufactured construction remains very unique in the complex balance

required to keep it on the competitive edge with customization without removing the

benefits of enhanced production with better quality, cost and time provided by

industrialization of process.











APPENDIX
REDUCED PROCESS MODEL: STROBOSCOPE'S LANGUAGE

/Control Statements; initialize resources content

INIT Orders 5 ORDER;
INIT Orders 2 ORDERB;
/INIT MatrlStrg 100;
/Ordering the Queue
DISCIPLINE Orders TYPE;
FILTER STARTORDER ORDER SimTime-TimeIn>2;
/Define Activities
COMBI Activate;
/PRIORITY CutFlrFRm 'Orders.CurrCount>OrdersB.CurrCount+l ? 10 : 0';
NORMAL Process;
/Define Links
LINK L1 Orders Activate;
DRAWWHERE L1 Activate.ORDER.TYPE.SumVal==Process.ORDER.TYPE.SumVal;


Process.ORDER.TYPE.SumVal==ORDERS.Type
LINK L2 Activate Process ORDER;
LINK L3 Process End;


DURATION Activate 2;
DURATION Process Normal[3,0.8];


/Run Simulation
SIMULATE;
/Run Simulation Until
SIMULATEUNTIL 'SimTime==200 I End.TotCount==7';
/Statistical Report
REPORT;
/Display additional Information
DISPLAY "Productivity is End.TotCount/SimTime houses per day";
/ASSIGN COLLECT PRINT









LIST OF REFERENCES


Abdelhamid, T. (2004). "Lean production paradigms in the housing industry."
Proceedings, NSF Housing Research Agenda Workshop, 12-14 Feb. 2004, Orlando, FL.
Vol. 2, 72-81.

AbuHammad, A., Hastak, M., and Syal, M. (2004). "Comparative Study of Manufactured
Housing Production Systems." Journal of Architectural F'iriit. rian'. 10(4), 136-
142.

AbouRizk, S., and Hajjar, D., (1998). "A Framework for Applying Simulation in
Construction." Canadian Journal of Civil Firiie riir'. 25(3), 604-617.

Archer, A. A. (2000). "A framework to integrate and analyze industry-wide information
for on-farm decision making in dairy cattle breeding." Ph.D. dissertation, McGill
University, Montreal, Canada.

Banerjee, D., Syal, M., and Hastak, M. (2006). "Material Flow-Based Facility Layout
Analysis of a Manufactured Housing Production Plant." Journal ofArchitectural
Eair1in t rinir,. 12(4), 196-206.

Barriga, Edgar M. (2003). "Manufactured housing industry: material flow and
management." Master thesis report, Purdue University. West Lafayette, IN.

Barriga, E. M., Jeong, J., Hastak, M., and Syal, M. (2005). "Material Control System for
the Manufactured Housing Industry." Journal of Management in Firi,'itc ri-'3. 21(2),
91-98.

Bashford, H. H. (2004). "The on-site housing factory: quantifying its characteristics."
(Jan. 1, 2008).

Center, B., and Verma, B. (1998). "Fuzzy Logic for Biological and Agricultural
Systems." Artificial Intelligence Review, 12(1), 213-225.

Chitla, V. K., and Abdelhamid, T. S. (2003). "Comparing process improvement
initiatives based on percent plan complete and labor utilization factors".
Proceedings, 11th Annual Conference for Lean Construction, 22-24 July 2003,
Blacksburg, Virginia, 118-131.

CityMyors Society (2003). "Affordable housing crisis casts a shadow over the American
dream." (Feb. 6, 2007).

Gan, B., Liu, L., and Jain, S. (2000). "Distributed supply chain simulation across
enterprise boundaries." Proceedings, 2000 Winter Simulation Conference, 10-13
Dec. 2000, Orlando, FL.









Hajjar, D., AbouRizk, S., and Mather, K. (1998). "Integrating neural networks with
special purpose simulation." Proceedings, 1998 Winter Simulation Conference, 13-
16 Dec. 1998, Washington D.C.

Hajjar, D., and AbouRizk, S. (1999). "SIMPHONY: An environment for building special
purpose construction simulation tools." Proceedings, 1999 Winter Simulation
Conference, 5-8 Dec. 1999, Phoenix, AZ.

Halpin, D.W., and L.S. Riggs. (1992). Planning and analysis of construction operations,
John Wiley & Sons, New York, NY.

Hastak, M., and Syal, M. (2004). "Building process optimization with supply chain
management in the manufactured housing industry." Proceedings, NSF-PATH
Housing Research Agenda Workshop, NSF and U.S. Department of HUD,
Washington, D.C.

Hastak, M., Jeong, J., and Syal, M. (2006). "Supply Chain Analysis and Modeling for the
Manufactured Housing Industry." Journal of Urban Planning and Development,
132(1), 1-9.

Jeong, J., Hastak, M., and Syal, M. (2006). "Supply Chain Simulation Modeling for the
Manufactured Housing Industry." Journal of Urban Planning and Development,
132(4), 217-225.

Jeong, J., Barriga, E. M., Hastak, M., and Syal, M. (2005). "Material Requirements
Planning for a Manufactured Housing Facility." Journal of Architectural
Eniin i. rini'. 11(3), 91-98.

Koskela, L. (1992). "Application of the new production philosophy to construction".
Technical Report # 72, Center for Integrated Facility Engineering, Department of
Civil Engineering, Stanford University, CA.

Koskela, L. and Howell, G., (2002). "The underlying theory of project management is
obsolete." Proceedings of the PMI Research Conference, Seattle, WA.

Manufactured Housing Industry (MHI). (2008). "Quick facts 2007".
(Jan.
3, 2008).

Martinez, J. C. (1996). "STROBOSCOPE: state and resource based simulation of
construction processes." Ph.D. dissertation, University of Michigan, Ann Arbor, MI.

Martinez, J. (2001). "EzStrobe-General purpose simulation system based on activity
cycle diagram." Proceedings, 2001 Winter Simulation Conference, 9-12 Dec. 2001,
Arlington, VA.









Mehrotra, N., Syal, M., and Hastak, M. (2005). "Manufactured Housing Production
Layout Design." Journal of Architectural Fi,'ii,, cri /,'. 11(1), 25-34.

Mullens, M. (2004). "Production flow and shop floor control."
(Jan. 1, 2008).

Narasimhan, R., and Jayaram, J. (1998). "Causal Linkages in Supply Chain Management:
An Exploratory Study of North American Manufacturing Firms." Decision Sciences,
29(3), 579-605.

Pietersma, D., Lacroix, R., and Wade, K. M. (1998). "A Framework for the Development
of Computerized Management and Control Systems for Use in Dairy Farming."
Journal of Dairy Science, 81(11), 2962-2972.

Senghore, O., Hastak, M., Abdelhamid, T., Abuhammad, A., and Syal, M. (2004).
"Production Process for Manufactured Housing." Journal of Construction
Engineering and Management, 130(5), 708-718.

Swaminathan, J., Smith, S., and Sadeh, N. (1998). "Modeling Supply Chain Dynamics: A
Multiagent Approach." Decision Sciences, 29(3), 607-632.

Thomas, D., and Griffin, P. (1996). "Coordinated Supply Chain Management." European
Journal of Operational Research, 94(1), 1-15.

Viswanadham, N. (2000). "Supply chain engineering and automation." Proceedings,
2000 IEEE International Conference on Robotics and Automation, April 2000, San
Fransisco, CA.

Wiezel, A. (2004). "Skill-driven optimization of construction operations."
(January 1, 2008).

Willenbrock, J. (1998). Residential design and construction, Prentice Hall, New Jersey.

Willenbrock, J. H. (2004). "The Leadership/Management Growth Model: Dynamic
Framework for Understanding Construction Management and Production in the
Housing Industry." < www.pathnet.org/si.asp?id=1068> (January 1, 2008)









BIOGRAPHICAL SKETCH

Diala H. Dandach was born in Beirut, Lebanon. In the year 2000, she graduated

from the Lebanese University with a Civil Engineering diploma with emphasis on

hydraulics. In 2003, she earned a DEA, equivalent of a Master of Science, in water and

environment from ESIB-USJ in Lebanon (Ecole Sup6rieure des Ing6nieurs de Beyrouth,

University Saint Joseph). She started working with a geotechnical engineering contractor

in Lebanon in 2002, where she spent 4 years working with a variety of responsibilities:

design, preparing proposals and bids, scheduling, budgeting and preparing contracts. She

also had the chance to work for a year on several projects that were starting in Dubai,

UAE where the company undertook new contracts. She then went on to pursue her

education and was accepted in June 2006 at the Rinker School of Building Construction

at the University of Florida where she was awarded the Master of Science in Building

Construction in May 2008.





PAGE 1

1 DEVELOPING STRATEGIES FOR CUSTOMIZED MANUFACTURED HOUSIN G By HAOURAA DIALA DANDACH A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN BUILDING CONSTRUCTION UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Haouraa Diala Dandach

PAGE 3

3 To my father, Rakan Dandach, and my mother, Sabat Daher, as an act of love and apprecia tion

PAGE 4

4 TABLE OF CONTENTS page LIST OF FIGURES.........................................................................................................................6 ABSTRACT.....................................................................................................................................7 CHAPTER 1 INTRODUCTION....................................................................................................................9 Overview of the Construction Industry...................................................................................9 Advantages and Disadvantages of Manufactured Housing...................................................10 Aim, Objective and Scope.....................................................................................................11 2 LITERATURE REVIEW.......................................................................................................13 Manufacturing Industries and Similarities with Manufactured Housing..............................13 Site-Built Construction and Differences with Manufactured Housing..................................14 Manufacturing Supply Chain.................................................................................................14 Lean Construction..................................................................................................................17 Improving the Efficiency of the Production Process in Manufactured Housing...................18 Improving the Efficiency the Material Flow and Management System in Ma nufactured Housing..............................................................................................................................19 Modeling and Simulation of the Manufactured Construction Process..................................20 3 METHODOLOGY.................................................................................................................23 Existing Mapping of the Production Process.........................................................................24 Customization of the Manufactured Construction Process...................................................26 The Management and Decision Making in a Customized Construction Process..................27 4 INCORPORATING CUSTOMIZATION AND MANAGEMENT STRATEGIE S INTO THE PRODUCTION PROCESS.................................................................................29 Customizing the Production Process.....................................................................................30 Managing the Manufactured Construction Process...............................................................33 Overview........................................................................................................................33 Defining the Decision.....................................................................................................34 Causes, Strategies and Outcomes of Decisions..............................................................35 5 CONSTRUCTION SIMULATION TOOLS AND FLEXIBILITY IN MODELI NG DYNAMIC DECISION MAKING........................................................................................42 Stroboscope’s General Characteristics for Dynamic Modeling............................................42 Strategies for Decision 1: “Do We Start Incoming Orders or Should We Wait ?”.........44 Stroboscope’s General Shortcomings in Simulating the Strategies...............................44

PAGE 5

5 Analysis and Modeling of Reduced Process.........................................................................45 Reduced Model: Simulation Language Deficiencies.....................................................46 Capacities of Adequate Simulation Tools......................................................................47 6 CONCLUSIONS....................................................................................................................50 Research Results....................................................................................................................50 Conclusions and Recommendations......................................................................................52 APPENDIX: REDUCED PROCESS MODEL: STROBOSCOPE’S LANGUAGE ....................54 LIST OF REFERENCES...............................................................................................................55 BIOGRAPHICAL SKETCH.........................................................................................................58

PAGE 6

6 LIST OF FIGURES Figure page 3-1 Methodology Flow Chart Representation..........................................................................24 4-1 Customized Production Process.........................................................................................31 4-2 EzStrobe Model: process map of customized orders, Order Type A................................32 4-3 EzStrobe Model: process map of customized orders, Order Type B.................................32 5-1 Simplified Customized Process.........................................................................................45 5-2 Flowchart for Requirements of Decision 1 Implementation..............................................48

PAGE 7

7 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science in Building Construction DEVELOPING STRATEGIES FOR CUSTOMIZED MANUFACTURED HOUSIN G By Diala Haouraa Dandach May 2008 Chair: Ian Flood Cochair: R. Raymond Issa Major: Building Construction Manufactured housing is assumed to be an improvement of the construction industry compared to the traditional site-built process. Industrialization of the construc tion process created many benefits including greater productivity, higher quality and r educed time and cost of construction. However, many obstacles exist facing the progress of manufact ured construction such as the inability to integrate more advanced technologies and automation in addit ion to the negative effect of customization on the production process: although customization of the process is feasible, including it in the process limits the benefits that might be achieved by industrialization, transforming the process into a labor intensive process. Simulat ion is one of the most effective tools to test management strategies and their effect on the pr oduction line in order to decide on the adequacy of using such strategies in the real-life process. T his research will analyze one of the most widely used existing simulation tools, Stroboscope, and its a bility to model customization and the dynamic decision making process in the manufactured housi ng industry. Guidelines of features and capacities to be incorporated into existing or new tools will be produced allowing these tools to model the stream of orders and dynamic decision st rategies

PAGE 8

8 and outcomes as a first step towards validating and testing these strategi es used in the manufactured housing construction process.

PAGE 9

9 CHAPTER 1 INTRODUCTION Industrialization of the construction process was introduced in the early part of the 20th century through the construction of manufactured housing. This innovation in the construction industry had many benefits over the standard on-site construction inc luding greater productivity, higher quality and reduced cost. In 1998, around 22% of new single-family housing was constituted of manufactured housing. This new industry manage d to provide the American home buyers with an affordable alternative offering qua lity, safety, cost-effectiveness and duration reduction. Overview of the Construction Industry In the past years, housing demand had risen considerably. In 1998, the housing demand was roughly short of 2 million housing units/year (Willenbrock 1998). Today, American cities are still suffering from shortage in affordable housi ng. Eighty per cent of the 1,000 large and small American cities surveyed by the National League for Cities in 2007 reported that rising housing costs are putting a severe strain on families. For example, Chicago (Population: 2.9m) identified an immediate need of at least 200,000 affordable units: in Minneapolis, Minnesota (Population: 383.000) over 50,000 units were identified and in Lodi, California (Population 67,000) identified a need of 8,000 units (City Mayors Society 2008). In face of the growing housing crisis which ma inly is an affordable housing crisis in the past two decades, the construction industry is fac ed with many challenges: the traditional site-built has been reluctant to us e innovation and technology to improve efficiency and maintain quality. This is due to its unique supply chain, site variability and the risks extending beyond the contractual liabili ty into the risk of reputation. Manufactured housing (MH) on the other hand is becoming more appealing

PAGE 10

10 to the American consumer especially since manufactured homes, unlike site-buil t homes, can benefit from mass purchasing of materials, products and appliances, adding s avings to the cost of purchase of the homebuyer. Advantages and Disadvantages of Manufactured Housing Manufactured housing is considered as one of the cheapest solutions at a rate of $40.80 per SF compared to a $91.99 per SF for site-built single houses in the year 2006 (MHI 2008). Approximately 65 corporations constitute the MH industry with 230 factories throughout the United States (MHI 2008). In addition to it being a cost-ef fective housing solution, MH housing has other advantages: for instance, a higher level of quality can be achieved since the construction is undertaken in a controlled environment including labor supervision, control of all aspects of the process and weather-free interference which prevents from delays. Better production rates and reduced construction time and cost are the results (MHI 2008). Despite this fact, the number of manufactured homes was noted to decline from approximately 350,000 units in 1999 to 117,000 units in 2006 (MHI 2008). This decline in industry can be reported to the fact that the MH industry is still primitive in its technology and modern equipment use compared to other manufacturing industries (Barriga 2003). Several constraints exist limiting the technology breakthrou gh of manufactured housing and mostly the fact that it is labor intensive and does not real ly incorporate many repetitions especially in the case of customization (Sengh ore et al. 2004). As a consequence, innovation and application of new devices to improve both the production process and material engineering areas become vital to the MH indust ry (AbuHammad et al. 2004).

PAGE 11

11 In an attempt to improve the efficiency of the process, different aspects have been researched and analyzed such as the production process, the material flow and its management. Approaches such as lean production and supply chain have been studied and adopted to generic manufacturing systems covering the process from the s uppliers to the customers. Attempts to provide better productivity were discussed to cover part s of the process and simulation was used to represent the processes mainly by networ k representation. Another main reason for the manufacturing industry decline would be the lack of customization: although possible and successfully adopted in the industrializ ed process, the customization posed a different kind of difficulties. The more customi zation and design options are available, the less the industrialized process becomes effic ient resulting in more overheads, coordination problems and idle time. However, throughout the literature review, very few papers were found mentioning the customization of manufacturing process and the development of generic models for decision making that will include customized orders. Aim, Objective and Scope This paper’s overall aim is to define and include customization in the generic representation of manufactured construction and represent the decision making proces s in the modeling of such a process. The objectives of this research are represented as follows: Identify the decisions being made and their components and define a generic decision making process representative of a manufactured housing construction process allowing for customization in the modules being built. Review of the existing construction simulation tools and their adaptability to represent the dynamic decision making process: Stroboscope, an advanced construction simulation language, will be analyzed for its efficiency and flexibility in representing such decisions and their outcomes.

PAGE 12

12 Suggestions concerning requirements from new construction simulation tools to better represent customization and the decision making process. The approach in this research is qualitative as to describing the decision making process and will be limited to finding of one site visit to a manufactured housing pla nt in Lake City, Florida. The construction process was found to include very few aspects of industrialization due to the management strategy used and to the few incoming order s: this created a labor intensive process and the productivity was not as sensitive t o decision strategies as expected. The main strategy was first in first out. In addit ion to the limitation of the case study findings, the research is also limited to only part of the constr uction process and does not include the overall supply chain.

PAGE 13

13 CHAPTER 2 LITERATURE REVIEW Manufacturing Industries and Similarities with Manufactured Housi ng In an attempt to study the possible approaches to better represent and analyze t he customized construction process, a review of existing approaches that have bee n applied to similar processes in other industries was performed. Complex systems we re approached in a variety of methods depending on the complexity, precision required and mainly the level of human interference. In cases where only qualitative approac hes desired, fuzzy logic has been used. This approach is related to the fact that comple xity and imprecision are two incompatible factors especially once the human intervent ion and analysis are involved: for instance, a fuzzy controller can be designed to repla ce a human operator in control systems. Complex systems are also approached with fuzzy lo gic like the evaluation of the performance of gas turbines (Center and Verma 1998). This met hod however describes the variables in qualitative rather than quantitative term s, which although applicable in areas like economics, bioengineering and traffic, can have s everal downfalls when used in the construction industry (Center and Verma 1998). A closer resemblance exists between the car manufacture industry and construction, mainly the manufactured housing construction where labor move from one product to the other in a fragmented process. Standardization, modularization have been used to simplify the process and increase productivity in an environment where there i s no variation in material flow, labor, production and demand. More importantly, customization seems to be the main barrier for the enhancement of the production proc ess in the project-oriented site built homes and the main reason of underachievement of manufactured housing (Wiesel 2004).

PAGE 14

14 Site-Built Construction and Differences with Manufactured Housi ng The site-built construction is the most popular construction method to date. However, this does not lead to conclude that it is the most efficient and productive construction method: in fact, it is in fact one of the most labor intensive and least productive processes; quality and productivity are difficult to control due to the exce ssive variability of the environment, the human factor (labor, management, subcontractors suppliers…) and many other variables that affect the overall process. In fact even when implementing the same project at close by locations using the same contrac tors, the cost, schedule and quality are never similar. In order to improve the site-built proce ss, some suggest a more modular framework for the structural and service systems inc luded in the construction process, separating the systems. This is expected to allow for m ore innovation and the introduction of technology in the actual construction process with the use of manufactured components. The goal would be a product with more interchangeable parts, lower cost, better quality, faster, higher quality a nd lower maintenance costs (Bashford 2004). It remains that more similarities exis t between the manufactured housing process and other manufactured industries than with the site-buil t construction process. Manufacturing Supply Chain Supply chain management (SCM) is the process of planning, implementing, and controlling the operations of the supply chain as efficiently as possible. Supply Chai n Management extends over all movement and storage of raw materials, inventory, an d finished goods from the origin to the consumption. Some researchers hypothesize that in general, a key factor to successful business is through managing the entire supply chain. Some researchers in the manufactured housing (MH) industry started focusing on

PAGE 15

15 studying the efficiency of the entire supply chain in comparison to the traditiona l manufacturing and quality based improvements. Coordinated planning of inventory-distribution integrated systems is the earlies t version of SCM and started as early as in the 60s. Research gradually progresse d to include first two or more stages of the supply chain and finally to models incorporati ng the three traditional stages in the supply chain: procurement, production and distributi on. Decoupling departments or functions (assembly, storage and distribution) of the same facility has proved to be an inefficient and non-competitive decision making policy (Thomas and Griffin 1996). A study of 215 North American manufacturing firms revealed the causal linkage s between sourcing decisions, manufacturing goals, customer responsiveness and manufacturing performance revealed the following: strategic sourcin g decisions (strategic outsourcing and supplier management) influences the degree of manufacturing g oal achievement (dependability, flexibility, cost and quality) which in its turn infl uences the level of customer responsiveness. However, the statistical tests show no str ong correlation between the manufacturing goal achievements positively infl uencing the degree of manufacturing performance. This can be explained by the fact that the labor productivity index used should be replaced by another index of manufacturing performance such as investment productivity (Narasimhan and Jayaram 1998). Consequently, simulation was researched as a tool for analysis and evaluation of supply chain design and management alternatives. Smith, Sadeh and Swaminathan (1998) base their framework on three main manufacturing industries: car manufactur e, computer manufacturer and grocery industry. Similar processes were identified desp ite the

PAGE 16

16 differences and two main categories were defined: structural eleme nts modeling production and transportation of products and control elements specifying various control policies. Supply chain interactions require more dynamic and sophisticated cont rols than the first in, first out (FIFO) queues. However, their framework and model basis di d not include continuous manufacturing simulation or decision-making controls modified wi th the evolving conditions (Swaminathan et al. 1998). Strategic alliances of independent (and international) companies create d the Integrated Supply chain Network (ISN). Modeling is a basis for effectiveness efficiency and optimizing cost, time and quality: strategic and capacity planning models a re nonlinear integer programming models whereas operational level decision ma king models use discrete event models (Viswanadham 2000). With the evolution of the SCM beyond the multi facility unique enterprise boundary into the cross-enterprise sc ope, technology was pushed a step further to allow for adequate cross enterprise si mulation that preserves confidential internal information of the concerned companies (Gan e t al. 2000). Supply Chain in Manufactured Housing: For years, SCM for MH industry focused only on the manufacturing and quality aspects of the process. Jeong, Hastak and Syal (2006) were the first to approach the SCM for MH as an entire supply chain using 5 manufacturers and 5 retailers in Indiana: the principle parties were defi ned to be the customers, the retailers, the manufacturers and the suppliers. For each group, the va rious characteristics and flows were detailed, analyzed and then grouped: the mat erial and information flow were combined showing all relations between the parties (Jeong et al. 2006).The conceptual supply chain model set the framework and basis for developing a

PAGE 17

17 simulation model in order to measure the performance of supply chain systems in pla ce or possible improvements of new alternatives (Hastak et al. 2006). The model is divided into four sub-models: house ordering, site and foundation preparation, house production in factory and house install and set-up. A dynamic, discrete and stochastic simula tion model ARENA was used and the following assumptions were made: double-wide house order, exponential probability distribution for customer generation and triangular distribution for the other processes, and the material availability is consiste nt. The performance measures selected were (1) how many house orders are gener ated in one year, (2) how long it takes a house from order to installation, and (3) the average que ue waiting time for each process. The main bottleneck was found to be from the time difference between the site foundation preparation time and the house manufact uring time: this leads to the conclusion that processing time at the factory needs to be improved. The scenario of a web-based ordering system was analyzed and found to optimize the SCM system (Hastak et al. 2006). Lean Construction By definition, lean construction is concerned with the “holistic pursuit of concurrent and continuous improvements in all dimensions of the built and natural environment: design, construction, activation, maintenance, salvaging, and recycl ing”. In construction, this approach attempts to improve construction processes by maximizi ng value and minimizing cost and accounting for cutomers’ needs. This accomplished through minimizing waste of materials, time, and effort in order to generate t he maximum possible amount of value (Koskela et al. 2002). Designing a production system to achieve the stated ends is only possible through the collaboration of all project

PAGE 18

18 participants (Owner, A/E, Constructors, Facility Managers, End-user) at early stages of the project, going beyong the traditional forms of contracts used in today’s industry. One of the applications of this principle to MH is the lean assembly where the process is simplified by industrialization, modularizations, standardization, a nd continuous flow processes, yielding to reduced waste and higher quality. The lean TFV (Transformation, Flow and Value) theory applied to construction management in the manufactured context defines it as the allocation of resource s transforming inputs into outputs while maximizing flow and value to the customer (Abdelhamid 2004). In a conventional construction management model, the critical performance measure is the capacity utilization. In a lean construction mode l, planning reliability is thought to increase system production. In fact, Chitla and Abde lhamid (2003) compared the conventional labor utilization factor (LUF) and the percent plan complete (PPC): the first one revealed problems like lost time due to measureme nts or travel whereas the latter reveled that the main issue was in the workflow bet ween stations in addition to the individual production station. Improvements suggested by the first included adding jigs and more inventory which will only create additional cost where as the latter suggested improvements included labor issues such as work pace, skill a nd education, material supply, unclear directions and other issues which, when addresse s correctly, will improve the throughput of the system (Chitla and Abdelhamid 2003). Improving the Efficiency of the Production Process in Manufactured H ousing Senghore (2004) analyzed the production process in the MH industry and defined it with five major areas: floors, walls, roofing, exterior finishes and interi or finishes. The process was however represented partly and does not include all the activities and stations from materials to finished product. In contradiction with the traditional

PAGE 19

19 construction management practices, it was found that improving the utilization of t he independent stations does not improve the overall production (Senghore et al. 2004). However, optimization of utilization of resources is still a common approach for many manufacturing companies and for researchers. To attain more fle xibility and better resource utilization, the facility layout design was studied in its present mo st common Ushape and was found to include much inefficiency (Mehrotra et al. 2005). Banerjee (2006) evaluated the layout design based on the material flow in a quantitative approac h based on optimizing material travel distances and costs (Banerjee et al. 2006). Some of the strategies for streamlining the production process concluded by Abu Hammad (2004) were to minimize the number of stations and create substations and split and merge activities so that the number of labor required will be consistent to finish work at ea ch station in addition to automating the movement from one station to another (Abu Hammad et al. 2004). Improving the Efficiency the Material Flow and Management System in Manufactured Housing The material flow and management system was not researched or studied with equal interest or attention until recently. This was due to the subjective scena rios in each factory transforming the production line into an efficient material manage ment process. According to (Barriga et al. 2005), the existing material flow and manageme nt systems consist of a material flow system and an inventory control system. The inventory c ontrol system allows for customized materials and in addition uses an independent demand system in opposition to the dependant system suggesting lean production and just-in-ti me inventory control systems. In any case, the generic material flow and m anagement system was developed by defining and incorporating all parties (suppliers, dealers and

PAGE 20

20 customers) and departments (purchasing, sales and production) involved. This gener ic system allowed for material requirement estimation and to relate dema nd for materials to the master production schedule noting however the drawbacks mentioned previously (Barriga et al. 2005). In addition, using a data base approach, an efficient material requirement pl anning system was developed managing the all data involved in material requirements estimation, allowing for better planning strategies and flexible with the c hanging demand (Jeong et al. 2005). Modeling and Simulation of the Manufactured Construction Process Static and floating bottlenecks along the production lines were found to be the main cause of delays along with other negative implications such as frustrat ion and exhaustion since workers pace have to continually vary (increasingly or decreas ingly) eventually affecting the quality of the product. This has affected the will of manufacturers to be innovative with new material and custom designs, affecting tremendously thei r competitive edge. The impact of bottlenecks is better evaluated by simulation: however, the variable management process restricts the accurate simulation (Mul lens 2004). The construction industry has not yet adopted simulation in its decision making process mainly due to the complexity of the process and to the extensive time require d to correctly simulate a process. Most applications still remain in the acade mic research level. General purpose simulation, usually presenting the user with more flexibil ity allowing for accurate representation, requires users to be knowledgeable in s imulation language. Some researchers argue that special purpose simulation addressi ng a specific sector of the industry will provide a user-friendly tool that will be adopted (AbouRizk and Hajjar 1998).

PAGE 21

21 Stochastic modeling is the common representation of variation. However this approach does not deal with the causes of the variation and could cause inaccuracy of the simulation results. The Artificial Neural Networks (ANN) approach is a too l capable of learning complex relationships between input and output data. The output of the ANN can be used as parameters to the simulation model. Such a generic integration model allows the external system to exchange information with the model for instance allowing factors like soils condition, soil classification, excavator and truck relative positioning to affect production in an excavation simulation model (Hajjar et al. 1998). In discrete event simulation, it is assumed that the state of a system chang es at specific times marked by certain event. Discrete event simulation is ver y adequate to model construction processes. These simulations can be performed using general purpos e simulation and special purpose simulation. Although general purpose simulations are more flexible since they can be manipulated to better suit the requirements of t he process, they require a lot of work and simulation knowledge which makes them less appealing in the industry. Special simulators and simulation languages are more domain-specifi c and have several approaches: Event scheduling (ES), activity scanning (AS) and pr ocess interaction (PI). Several simulation systems use diagram based on activities or activity cy cle diagrams (ACD) represented by networks consisting of nodes (activities a nd queues) connected with arcs or links. These systems clock advance and apply both the AS strategy combined with ES for event generation mechanisms. Many example s of simulation tools or programs exist such as: Stroboscope, Cyclone cyclic operati on network, Hocus hand or computer universal simulator and many others (Martinez 1996).

PAGE 22

22 Stroboscope is a very advanced simulation language/tool capable of modeling very complex systems. However, it remains very difficult to use since it req uires knowledge of the language used. EzStrobe (Martinez 2001) was developed to facilitat e the use of simulation tools with a user-friendly modeling graphical interface. It uses the basic Cyclone ACDs, Stroboscope simulation software and the MS-Visio graphical interface. It also allows to some of Stroboscope features such as the use of vari ables to parameterize the construction process and allows for standard and customized re sults. However, it falls short of Stroboscope’s flexibility and ability to program ver y complex interactions especially the features relating to resources and their types: resources in EzStrobe are not differentiable when residing in the same queue. However, it i s still useful for the purpose of applied research.

PAGE 23

23 CHAPTER 3 METHODOLOGY The approach to this study is qualitative in defining the problem and suggesting alternative solutions. Initially, the existing production process will be descri bed in detail with all its components. In a second step, customization will be added to the existing simulation models in attempt to represent the flexibility in options that the industr y offers. This will be done using EzStrobe, a visual and simplified version of Stroboscope used in a previous research to represent the production process (Senghore et al. 2004). Consequently and in a more detailed approach of the customization, specifically the decision making process used in most MH management techniques, a detailed analysi s of the decision making process, components and strategies will be performed simultaneously with a comparison to existing simulation tools and languages and their adaptability to model such decisions and strategies. Given that it is one of the most widely available construction simulation tools, Stroboscope will be mainly anal yzed for this purpose. Guidelines for developing existing tools or creating more flexible t ools that will better represent the dynamic construction decision making process will be presented as the result of this research. The above mentioned methodology sequence is summari zed and represented in the following Figure 3-1.

PAGE 24

24 Figure 3-1. Methodology flow chart representation Existing Mapping of the Production Process As stated in Chapter 2 describing the literature review for this study, res earch studies analyzing the MH production process exist: these studies reduced the proces s to a model used for comparing and analyzing factors such as labor and material utili zation in

PAGE 25

25 an attempt to improve the process. The process was mapped through case studies and was assumed generic to most manufactured housing processes. In general, the process wa s represented similarly to the physical findings at the MH plants: by types o f works (or activities) performed at one or more physical stations. The main five types of works were (1) floors, (2) walls, (3) roofing, (4) exterior finishes, and (5) interior finishe s. A plant can have many stations (up to 16 stations depending on available space in the plant, on the complexity of the homes and on the planning of activities in the various stations In parallel to the main production line, various feeder stations support the various activi ties with resources like cabinets, fixtures and roofing supports (Senghore et al. 2004). E ach station was mapped separately showing the logic of production flow, the activitie s and processes used including durations and logical sequencing, and required resources associated with the activities including availability and quantities used (labo r and material) in addition to the stations’ usage. This mapping was used to represent the process through simulation in order to test the efficiency of the stations, resource s and determine the causes of delaying the process if any in order to improve productivi ty. Limitations: However, this approach is limited by various factors resulting in an inaccurate representation of the actual process. On the one hand, the research studie s only cover bits and pieces of the manufactured housing process: studies were limited to some areas of the constructions (wall frames to exterior finishes) and did not include the remainder of the process till the finalizing of the house. In addition, the areas i ncluded in the model would not be entirely represented and broken into the detailed activities but it was limited to a more generic approach. This is mainly due to the complexity of approaching the subject as a whole. However, this resulted in a restricted analy sis of the

PAGE 26

26 system that does not truly represent the real life system. On the other hand, the production process was approached in the representations but not the material and resource flow and restrictions: in the actual MH plant, resources including labor crews and materials are managed in a more flexible manner and are conditioned by the de cision of the plant production manager; in addition, the process does not account for the diversity in customized housing and the various types of models that requires different amounts of resources and occupies stations and activities over different durations. As a result, since the representation of the process is not dynamic enough to represent t he flexibility and variability of the real process, the analysis and optimizat ion of the current modeled process cannot be considered best representative of the optimization of the r eal system. In fact, the majority of the tools used in the literature for the simula tion of MH are not flexible enough to include the dynamics of the process, the customization and the conditional decision making. Customization of the Manufactured Construction Process It was found that the customization once incorporated in the construction process in MH has created many inefficiencies in the process and the productivity of the system: the more the customization in design and construction, the less the industrialized proces s becomes efficient resulting in more overheads, coordination problems and idle time However, the existing research did recognize that customization in the MH i ndustry is a requirement vital for its competing end. However, it has not been included in the models as a major player in the optimization and output of the process: incoming orders were considered as a continuous flow of similar models that will not affect the sequenci ng of processing, timing of starting activity processing, resource usage or act ivity durations. In reality, orders are not processed on a first come first serve basis rega rdless of their types,

PAGE 27

27 numbers and the types and numbers of orders already being processed in the production line. Decision making remains a subjective process that relies mainly on the MH plant managers, their personal experiences and judgment. However, the process being as complex and dynamic as it is, the need for introduction information technology specifically computer simulation software that are tailored for this purpose ha s become critical especially sine the MH industry remains inefficient when com pared to other manufactured industries such as car industries that have managed indeed to include customization in the optimization of their processes. The Management and Decision Making in a Customized Construction Process For the purpose of this research, a case study of a manufactured plant in Lake City Florida was considered. This plant produces a diversity of models and is very flexi ble for customization beyond its multitude of model plans. The process used in the plant was considered only as basis for this research and analysis for the following rea sons: The overly customized process has reduced many of the industrial manufacturing benefits. In fact, over-customization has left little common components being manufactured that the process resembled the site-built except for the controlled environment: the process was found to be very labor intensive. Orders have not been as abundant as expected especially with the housing crisis taking place. As a result, large delays between orders have resulte d in a reduction in labor and space resources: many crews were laid off and only one of the many plants in functioning. This also relieved the resources and stations from the expected constraints in the production line and the productivity or output (mainly the number of days per houses) was mainly affected by the frequency of the incoming orders and not the efficiency of the resource allocation and process management or the productivity and usage of the resources at the stations. As discussed, the case study will only be used as a basis for qualifying the poss ible decisions to be made by the plant manager in the case of various customized orders incoming at different time intervals. These decisions will be described in deta il as far as

PAGE 28

28 all the factors affecting them, the strategy and conditions that decide on the decision. Stroboscope (Martinez 1996) is described to be the most flexible and adapted construction simulation tool, providing its users with a dynamic simulation languag e instead of a restricted template tool. This language’s improvements over other sim ulation tools mainly lie in the fact that it can dynamically and actively consider t he state of the simulation. This language will be discussed and analyzed to see its adequacy to repre sent and incorporate the decision making process into the production process of any MH process. Future recommendations will be made based on the deficiencies of the exist ing construction simulation tools and how they could be overcome.

PAGE 29

29 CHAPTER 4 INCORPORATING CUSTOMIZATION AND MANAGEMENT STRATEGIES INTO THE PRODUCTION PROCESS The manufactured housing production process has been extensively discussed and analyzed as shown in Chapter 2 that describes the literature research. The proc ess has been described in its entirety starting from the suppliers to the homeowners and wa s modeled in partitions due to the complexity of representation. The customization of the manufactured housing process although recognized as vital for attracting homeow ners has not yet been fully integrated into process mapping or any simulation model encountered in the literature research. Another aspect that was totally igno red was the plant manager strategies and dynamic decisions: the orders, in addition to being considered of only one type, were also processes as soon as they arrived. The only restrictions on the production line were the availability of the resources, the available openings in a station without any consideration to any strategy as far as st arting the process (lining up similar orders for instance) or managing the production line (m oving around resources or stopping one station’s activities in order to allow other orders to be processed). As part of the objectives of this research, we will first approach the customiz ation of the incoming orders and include it in the model of the production process. This will be performed using EzStrobe, a graphical simplified simulation tool that uses St roboscope’s simulation language and Visio as a graphical interface. This simulation tool ha d been successfully used to model part of the production process by Senghore (2004). In a second more detailed approach, a series of management decisions of the production line were defined and detailed showing all the factors affecting such decisions, conditions for evaluations and outcomes of the strategies used. These decisions

PAGE 30

30 were an estimation of possible decisions that might affect the production of the pla nt and were based on observation from site visits and the literature. After defining t he components of the strategies, Stroboscope, being the most advanced construction simulation language, was analyzed to test its flexibility to model the decisi on making dynamic process: flexible aspects as well as its shortcomings were d escribed in order to draw conclusions as to what would the required capacities of new construction simulat ion tools be. Customizing the Production Process In a first attempt to incorporate customization into the process and the nonpredictable incoming order variability, and based on findings during a visit to the previously mentioned manufactured housing plant, a reduced process map was defined to include several activities including (1) flooring framing, (2) wall frami ng, (3) electrical and plumbing, (4) insulation and (5) drywall. The available stations were considered as a resource in addition to labor (construction crews, carpenter crews, electrici ans and plumbers’ crews, insulation crews) and material for carpentry. EzStrobe (M artinez 2001) is a simulation tool developed as a user friendly extension of the simulation langua ge Stroboscope which conserves some of its features including the following: The use of Cyclone’s basic modeling features including queues, combi activities and normal activities; The use of stroboscope as the simulation software while the user can graphically use a template using MS-Visio graphical interface; and The use of parameters and the generation of customized results. In a first approach to the customization issue, EzStrobe was used to model and simulate the above described process map as shown on Figure 4-1.

PAGE 31

31 Figure 4-1. Customized production process The main activities (all of them combis) are cutting of the floor framing, c utting of the wall framing, construction of the floor framing, construction of the wall fr aming, Electrical and plumbing works, insulation works and installation in drywalls. The following constitutes part of the logic and conditional sequencing of the process: The first two activities require only start with the incoming orders but are considered parallel to the production line since they do not use the station resources. Floor framing is conditioned only by the activity for cutting the floor frames and can start independently of cutting of the wall framing which conditions the wall framing is not ready. A continuous material generator is added in order to prevent any shortage from material to affect the process; this, however, is not representative of the real life scenario where at times material falls short and does creat e some delay in the process. Carpenters and materials are shared by the cutting of frames activi ties, with a priority given to the cutting of the floor framing. All other activities require labor crews and the preceding activities to be executed. A station space will be occupied when floor framing starts and will only be available after the drywall is installed.

PAGE 32

32 Two types of orders are included in this model sharing the same labor, material and space resources as shown on Figure 4-2 and Figure 4-3; however, they included different amounts of usage of the resources and different activity durations. In addition, the two order types are generated through probability links shown on Figure 4-2. Figure 4-2. EzStrobe model: process map of customized orders, order type A Figure 4-3. EzStrobe model: process map of customized orders, order type B

PAGE 33

33 The model performed the task allocated in representing the customization and sharing resources: with the roughly assumed values, an output of around 2.5 days per house was established. However, the use of EzStrobe was overruled because it did not provide the flexibility required conditioning the sequence and the logic behind the customization and it was clear that it will not be resourceful for integratin g dynamic decision making and characterizing resources. As a result, although the interface was very easily handled, it did not provide any of the features that are assumed available with Stroboscope and that part of the rese arch was only included to rule out the use of EzStrobe. The research was then diverted to defining the decision making process and all the factors relating and affecti ng it in addition to the conditional strategies and their outputs. Consequent to defining them, the use of Stroboscope will be qualitatively assessed to check its suitability to r epresent the resources, the strategies and the overall decision making process. Managing the Manufactured Construction Process Overview All manufactured housing plants have a production manager or a plant manager that actually makes decisions concerning the production and material manageme nt. Material management was analyzed in previous research: a material supply management system was defined that was used in improving the material requirement es timation process (Barriga 2003). Activities were only considered in the context where they influence the material flow and management system. The system uses backwa rd scheduling and introduces inventory control and supply chain management concepts to the MH industry. However, this study did not include the economic implications through the efficiency of the production system which still needs to be tested. In fact processing

PAGE 34

34 orders, especially when they are customized orders, on a delivery time basis onl y is probably most cost efficient as far as the production throughput and resources usage are concerned. During the site visit to the plant, it was noted that no specific decision making process was in place and that was mainly due to the low rate of ordering of houses In fact, one of the noted strategies was to randomly alternate types of ordered home s in order to create diversity for the labor force which is supposed to be an incentive f or productivity. However, this strategy was not tested to prove its efficiency. R egardless the reasons why such strategies are not in place, whether it is the decreasing amount of orders due to the construction crisis or the lack of testing and control tools that allow such decision to be virtually tested prior to their application, this remains a new field t hat have not yet been researched in the MH industry. Defining the Decision Since no information was found in the literature or from the plant visit, several scenarios of decisions were determined using the best judgment and common sense. Af ter defining a series of decisions, the factors affecting those decisions we re mainly defined from 3 types of variables: the variability of type and number of the incoming orders, t he present state of the process activities (expected ending time of activiti es) and resources (usage and availability). The following decisions were defined as guidelines for possible logical str ategies to be tested for their efficiency as far as the throughput of the production line (number of days per house for instance): 1. Starting incoming orders or holding them.

PAGE 35

35 2. Stopping processed orders and moving them out of production line or letting them be processed. 3. Moving resources around from one station to another. 4. Produce inventory and store it. The above defined decisions are independent headlines that cover various aspect of the plant management. They require strategies to be determined and implement ed once the conditions for implementation are established. The outcome of the strategies ne eds to be also defined as to what aspects of the process will be affected. Each decision w ill be considered and analyzed independently from the others. Once these strategies a re defined, they will have to be tested at all times during the process and this will require, in the case of a simulation, the need to access many characteristic of the simul ation at all times. And this precisely is the main obstacle faced in the common simulation tools Causes, Strategies and Outcomes of Decisions Each of the enumerated decisions is analyzed separately in the following sub paragraphs. The first decision will be analyzed and a strategy will be conclu ded for its implementation and the modeling of such a strategy using Stroboscope will be discuss ed in a subsequent chapter. The second decision will also be thoroughly analyzed; however, due to its complexity and the new constraints it introduces, a strategy will not be concluded. The other decisions will be moderately detailed as to the factors rela ting to their implementation without going into further details. Decision 1 The first decision can be phrased in a question form: Do we start incoming orders or should we wait? In other terms, when should we start processing the incoming orders?

PAGE 36

36 The factors or variables affecting the decision depend on the distribution of the incoming orders and the current state of the orders being processed. For the incoming orders, some of the data required are the number of orders, their time of arrival and finishing deadlines if any. For the orders being processed at the time of the incoming order, the required data are how many orders are being processed, what are thei r types and their expected finishing time. However, the plant is not considered one block where once the order is started it will be processed to the end. On the contrary, the production line is composed of a series of station where a variety of independent activities are being performed. These activities require their own resources and only share the physical space or station with other activities. In order to have access to the expected finishi ng times of any order, a manager has to have access to information such as the state of each a ctivity: the percentage of performed work and the remaining duration in case the activity started and expected duration and availability of resources in case the activity is rea dy to start. A previous study of the production process showed that, in contradiction to traditional mass production paradigm, decreasing labor utilization (by adding resources to mi nimize activity duration) leads to a better over all production (Senghore et al. 2004). This may lead to assume that creating activity sets that require similar processi ng durations, regardless of the number of resources needed to create that situation will probably result in a better overall productivity of the process. Consequently, access to the remaini ng duration for activity completion is necessary for the decision making process in t his case. This may also lead us to consider processing similar order consecutively for a be tter efficiency of the production line: comparing the state of the production line and the incoming orders, if the incoming order is of the same type, we might process it as soon as

PAGE 37

37 an available station (in addition to material and resources) are available; i f it is of a different type, we might choose to not process that order and maybe wait to have a few similar orders before allowing them to occupy the production line. Waiting time restriction will be required not to exceed a certain limit if delay of a cer tain home will create additional cost that exceeds any saving in the improvement of the productivit y of the process. After determining the factors that affect this decision and the required data from the system, the following strategy was developed in order to be eventually teste d: If incoming orders are similar to the ones being processed, then we process the order as soon as we can. If incoming orders are different then ones being processed and ordering history or prediction shows there will be other similar types to that order, then we hold the order with a different type and we process it with the similar ones. Do not hold orders depending on their types for more than a specified period of time (that can be tested). An example simplifying this strategy would be considering the production proc ess described on Figure 4-1 and considering two types of homes to be processed (type A and type B): if the activities on the production line are processing all type A m odels and a type B model is incoming, then we hold that type and only process new type A models. That only happens until a point in simulation time where the type B is delayed beyond an acceptable time and it requires being processed. At that point, we start process ing all type B models held and incoming up to the point where type A models A held are delayed beyond the acceptable deadlines.

PAGE 38

38 Decision 2 The second decision can be phrased in a question form: Do we stop orders being processed and remove them from the production line? Or in other terms, when should we stop orders being processed in order to allow incoming orders to start being proces sed? Similarly to the first decision, the factors or variables affecting thi s decision depend on the incoming order type, number and distribution in addition to the state of the processed orders (type, numbers and activities’ remaining durations). With the same logic used with the first decision, if one processed order (or more) is creating a dela y in the production line by taking longer time for performing the activities at the diffe rent stations which holds up the stations downstream and creates idle time for the station upstream then a decision must be made about whether to take that order out of the production line, add resources to push it through the production line or just leave it on the production line. This decision should be compared to the alternative of replacing this order by another incoming order of similar type to the other orders being processed (which means with similar activity processing durations). However, the decision is to stop the acti vities for this order and remove it from the production line, the scenario tends to be more complex than in the case of the first decision. The first complexity lies in the fact tha t access to the state of an activity although easy but the destruction of such an activity due to a dec ision process is not that easy to implement: in the most recent simulation model, once star ted, an activity will be processed till its end before any condition is checked. This cr eates a first obstacle in modeling such a decision with the most recent and common construction simulation tools. If this was overcome, the second obstacle lies in how to release t he resources held by that activity before its termination in addition to the third obst acle concerning how to reintroduce this order back into the production line and under what

PAGE 39

39 conditions. The data required by this type of decisions is far more complex to obtain t han that of the first decision since it involves termination of activities being exe cuted, release of resources before activities are performed and introducing partly execut ed activities. As a simplifying example, we can consider the production line described in Figure 4-1 and the same conditions described for the decision 1 (type A and B models). Assuming the two type B, one type A and two type B models being processed in that order and considering that model A takes a longer time to be processed, we can consider removi ng the type A model from the production line and replace it with an incoming type B models. Decision 3 The third decision can be phrased in a question form: in the case where some activities are delaying the production line, should we move resources from one st ation to the other? Or in other words, do we allow flexibility in moving resources in order to push activities delaying the production line and keep the flow of production? This decisi on concerns resources but it is conditioned by the status of the production line and the activities being performed. When stations start delaying the production line because they are taking longer than the other stations, a good strategy might be to accelerate the activitie s in that particular station. This can be accomplished by adding to the labor resource which will accele rate the performance of the activities. As discussed in the second decision, with all the impracticality of implementing it into a simulation model to be tested, the othe r strategy would be to take that model out of the production line. Similarly to decision 2, the strategy of moving resources around is not very practical to implement into a sim ulation model in order to be tested. In most construction simulation tools, the only condition on

PAGE 40

40 resources is their availability and required amount in order to allow activitie s to start. Once the activity starts it holds the resources and will not release them unti l the end. In addition, the simulation tools do not allow activities to accept resources during the performance of the activity or in other term before the activity is complete ly executed. Conditional allocation of resources in the case of a slow activity seems impossi ble with the existing construction simulation tools and further research to approach this ma tter should be directed toward innovation in simulating such a strategy. Decision 4 The fourth decision is not precisely a decision since it represents a performanc e strategy that will be either adopted in a plant or not and does not require a dynamic decision making process and changing conditions. However, we will still include t his in this section since it will require similar strategies, inputs and outputs reg ardless of the fact that it is not a dynamic aspect that changes along the production line. This performance question concerns the choice of keeping inventory: will keeping inventory increase the production flow and the overall throughput of the system by reducing any delay that might result from lack of material inventory? Is this increase in productivity considered profitable when comparing the increase in production (if any ) to the cost of keeping such an inventory (cost of providing space, resources required to move material, tying capital in inventories that will wait a certain peri od of time before being processed). This strategy does not seem hard to model and test with the traditional construction simulation tools. However, if we consider inventory to comprise not only raw mater ial but also modular components that could be produced for the various types of homes, the problem starts to be complicated: preparing such modular components will require a

PAGE 41

41 labor force that will not be always required to perform on the inventory activities. Suc h labor crews can be considered an additional backup for the existing crews. In this case, limitations and conditions on when to stop production modular components and move the labor to help the other crews on the production line falls into the complexity encounter ed in modeling the strategy of the third decision (moving resources around when certa in stations are delaying the production line). In the following chapter, only the strategy of the first decision will be addres sed and its main modeling components and difficulties will be identified. As mentioned in a previous chapter, Stroboscope (Martinez 1996) is supposed to provide its users with an innovative dynamic simulation language. For the purpose of this research, the flex ibility of this language to model those difficulties will be discussed in a qualitative manner providing feedback whether such a language is adequate to solve this issue or other resources will be needed.

PAGE 42

42 CHAPTER 5 CONSTRUCTION SIMULATION TOOLS AND FLEXIBILITY IN MODELIN G DYNAMIC DECISION MAKING The following chapter will consider the first decision described in Chapter 4 and analyze the possibility of modeling it along with the customized production process using Stroboscope. As mentioned previously, out of the widely available simulation tools for modeling construction processes, Stroboscope is the most used and most flexible as thought by its authors. It is thought to be the most flexible in modeling the dynamic features of the manufactured construction processes. Stroboscope’s main feature that will be used is its characterized resources: this feature will be used to represe nt some of the complex conditional resource release taking place during the decision making pro cess. On the other hand, other aspects of the decision making process that cannot be solved using the features of Stroboscope will be described and possible other modeling solutions will be suggested. Stroboscope’s General Characteristics for Dynamic Modeling The customization process described in chapter 4 and represented using EzStrobe in Figure 4-2 and Figure 4-3 does not efficiently represent the customized product ion process. This is mainly due to the representations of the stations as resources: the stations are not all similar and the ordering of the stations, which is not represented in the EzStrobe model, is very important. In fact, having two similar activities s equence sharing the same resource is not the best representation of how the production line processes two types of orders: in the previously represented model, if enough resources are av ailable, the station for floor framing for instance could be occupied by two models or more bef ore the station resource moves to the next activity; this is not possible in the real lif e scenario. In addition, this representation forces the sequence for all activities and does not a llow

PAGE 43

43 for flexibility of independent management of the stations: once an order is sta rted, it will be processed till the last activity which does not reflect the flexibility of the real scenario where decisions at each station can be made independently. Stroboscope’s most important feature relating to our subject matter is the simulation of unique “characterized resources” in opposition to “generic resources ” which are interchangeable and indistinguishable. Characterized resources r epresent more accurately the real life uniquely identifiable resources. In addition to fix ed properties that identify and characterize certain resources, characterized resources us ed in Stroboscope have other properties called “SaveProps” that change as the simulation is r unning and are defined by the user according to their need. Other properties of the characterize d resources are system maintained and are very useful for identifying resourc es: BirthTime which is the value of the simulation clock when at the time of creation of resource ResNum which is a serial number unique among resources of the same type TimeIn which contrarily to the other two is not fixed; it represents the value of t he simulation time when the resource entered its previous node; it is updated every time a resource enters an activity or a queue. Other properties for resources exist such as VarProp which create the abil ity of creating properties that are functions of other resource properties. Strobosc ope also allows for access to system maintained statistical variables of resour ce properties in both queues and activities. These properties are only available when nodes (activiti es and resources) are cursored: in the case of resources in activities, acces s to properties is only possible when activities are either starting or ending. In any case, in the fol lowing, we will discuss the flexibility of stroboscope to model the decision making process in a customized manufacturing scenario.

PAGE 44

44 Strategies for Decision 1: “Do We Start Incoming Orders or Should We Wait?” The strategy for this decision was to process similar types of orders in a s equence in an attempt to improve productivity. If an incoming order is similar to the ones being processed, then this order will be processed as soon as possible when the production line allows it. If the incoming order is different than the ones being processed, this or der is held to allow for other orders to be processed. This however is limited by two variable s: the first restraint is a time limitation where an order should not be held for longe r than a certain period of time (d days). The second constraint is a limitation on the number of orders allowed to be held: after a certain number (n) of held orders of a certain t ype, those orders can then be processed in a sequence. Stroboscope’s General Shortcomings in Simulating the Strategies In order to achieve this logic, access to certain data is required: Orders should be characterized and differentiated as to their type, the time and date of receipt and their waiting time before being processed. The orders being processed at the various stations (and activities): access to the type of the order resource that is activating a certain activity (combi ). Stroboscope allows for several attributes for characterized resources tha t allow for a determined and flexible resource acquisition for instantiating activities through drawing resources from queues; these attributes include: ordering of the resources i n a queue according to a specified “discipline”, an order drawing through the draw links and s orting when drawing (DrawOrder), drawing under certain conditions from a queue (DrawWhere) and drawing until a certain condition will be met either in the queue or i n the instantiated combi (DrawUntil). One of the benefits of DrawOrder is that it can regulate drawing of resources for two different activities requiring dif ferent sets of conditions. However, such a sorting affects greatly the simulation time.

PAGE 45

45 Similarly to its drawing feature, release features of the charact erized resources have the same behavior: the same type of conditional release applies to the release of resources from the activities that are terminating using attributes such as Relea seOrder, ReleaseWhere and ReleaseUntil. Analysis and Modeling of Reduced Process In an attempt to test Stroboscope’s flexibility to model the first decision sc enario, the problem was broken down into its simplest aspect in an attempt to simplify it for the purpose of this study. Along the same logic describing the production process in the previous chapters, resources represented by two different types of orders ar e considered along with only two activities. The resources are initiated as five resources of type A and two resources of type B. Order resources are qualified by their type and pro cessing durations. The process is represented on Figure 5-1. Figure 5-1. Simplified customized process The different conditions for drawing orders were defined according to the types and count of the resources. The conditions are detailed as follows: As a first general rule, when the order type in “Activate” and “Process” are similar, then draw the order with the same type from the queue Orders. If the orders in “Activate” and “Process” are not similar, then check the number of orders of each type in the queue: Start processing the one with the largest number. Never keep an order for longer than 2 days without being processed.

PAGE 46

46 The first obstacle was in the fact that duration of activities is not flexible de pending on the resource type: The duration pf processing one type is different than the duration of processing another; hence duration of the activity and the type of the order should be connected. Reduced Model: Simulation Language Deficiencies For the purpose of this paper, the main features of Stroboscope were studied in an attempt to capture its main abilities to represent and simulate the customiz ation in the MH process and the decision making process. It is to be noted that due to many restriction, this study acknowledges its limitations as to the full understanding of the details of the simulation tool at hand: this is due to the time constraint along with the complexity of the language itself that requires deep and lengthy evaluati on. In acknowledging this fact, we highlight the possibility of overlooking and failing to us e some possible features that could have been used to work around obstacles found in representing the researched process with Stroboscope. The simplified customized process represented in Figure 5-1 was represente d using stroboscope. The difficulties were encountered in modeling the logic of the strat egies enumerated in the preceding paragraph where the use of the features was not as flexible as expected and lead to errors and misrepresentations. The first difficulty encountered was in the assumption that we can at least or der the resources by their waiting times. However, this is not possible since using the “Discipline” function, it will evaluate a resource only one at the time it ent ers the queue and will not reevaluate to check if this resource has been waiting for 2 days (as required). In order to compensate, the function “filter” was used; the filter forms a subs et of resources within the queue but it still did not solve the issue.

PAGE 47

47 Another related matter to this one is in the generation of various distributions of the same resource of different types: although, using “Generate” allows for generation of resources during the time when the simulation is running, this creation of resources i s on the one hand restricted by the start and end of events (such as start or termination of an activity) and on the other hand such a generation cannot be randomized in the case of Stroboscope’s characterized resources (it could be in the case of generic resour ces). When trying to compare the type of resources contained in the activities, it was impossible to do so in Stroboscope: using “DrawWhere” which allows selecting resources that satisfy certain conditions; this is due to the inability to compar e resources that are in activities in the case where the resources are not cursored. In fact, the use of this function, in such a case, could lead to a run time error which interrupted the simulation. The software is attached in Appendix A to show the attempts made to run the simulation. However, the overall impression was the inability of Stroboscope to simul ate the complex and dynamic decision making process even in a simplified process like t he previous one. Restraints are due to the aspects of the decision making process that ne ed access to certain information about the resources such as the types of models being processed and what type of these resources are being processed. Capacities of Adequate Simulation Tools Based on the previous discussion regarding the deficiencies and limitations of t he existing construction simulation tools, and based on the requirements of the manufacturing plant for decision making strategies and implementation, a basi c flowchart for the conditions, processes and required input data for the implementation of Decision 1

PAGE 48

48 was represented as shown on Figure 5-2. The incoming orders are basically chec ked for three main conditional decisions prioritized in the following sequence: The first condition refers to resource delays in the queue holding the orders: if orders have been delayed for more than a specified allowed value, then these delayed orders will be processed. If all delays are still within l imits of time delay, then orders can still be held in queue and other conditions for processing will be checked. The second condition refers to activities in the production line and the orders that they are processing: if they are processing orders similar to an y of the incoming orders, then this type of order will be processed; if not, the third condition will be checked. In case none of the previous conditions were fulfilled, a third condition for processing should be checked: independently of time, orders held and not processed should not exceed a certain specified number. When incoming orders held of a certain type exceed this number, orders of that type should be processed by the order of their incoming time. Figure 5-2. Flowchart for requirements of decision 1 implementation

PAGE 49

49 These strategies for decision require access of specific data concerning activities and resources at all times and not just during the time where an order is incoming or an activity is being processed (activated). Such a language or tool should give acc ess to statistical data cross-referenced with the simulation time. The above mentioned logic is intended as a basic guideline for requirements of a new simulation tool or language or adjustments to be performed on existing simulation tools such as Stroboscope. The strategies will require an adequate simulating t ool that will effectively represent the conditions and make the required data availabl e. Once properly modeled, these strategies should be tested as to their efficiency in i mproving the production. Other strategies will also be represented, simulated and compared and the best strategies will be used for real life implementation to improve plant product ivity.

PAGE 50

50 CHAPTER 6 CONCLUSIONS This study provided a basic analysis of the current state of the manufactured construction industry in referenced to customization. After a brief analysis of the benefits of such customization and based on the literature, a set of decisions and their respecti ve strategies were defined to incorporate customization and to be simulated and test ed. In a parallel manner, a more detailed analysis on existing simulation tools was per formed to analyze their flexibility in simulating such strategies. As a result, the deficiencies of the existing construction simulation tools were established and recommendations for improving those tools to better represent the construction process were suggested i n an attempt to ameliorate and upgrade the simulation process and eventually the eff iciency of the testing strategies in order to implement them in real-life. This chapte r will include the results of this study and the conclusions and recommendations provided for future research. Research Results Following the sequence defined in Chapter 3 describing the methodology and steps followed, various results were established at each step of this study. First, d uring the preliminary research of the construction manufacturing industry, the importance of customization was established as the most defining element for the competitive edge of the industry in comparison with the rather unique sit-built construction. However, allowing for too much customization, although easily applied in manufactured construction, will produce many obstacles to the industrialization process itself creating a labor intensive process that no longer benefits from automation of the process or ot her benefits of industrialization. On the other hand, and throughout the literature reviewed

PAGE 51

51 during the course of this research, it was found that the production process in manufactured housing was analyzed and modeled independently of the customization: in fact, most of the studied reviewed did not include or allow any customization in the process and only one type of homes was being processed. Consequently, all the aspects of management and decision making were not incorporated in those studies. This repres ents on of the major holdups that the manufactured housing industry has been suffering from, compared to other industries such as car manufacturing where automation and simula tion have been largely introduced and used to improve the processes. These results lead to the second part of this study where a set of dynamic dec isions and strategies were defined and detailed to represent possible scenarios that need to be introduced in modeling and simulating the production process in a manufactured construction plant. Concurrently, an analysis of the existing tools and their flexi bility to model such dynamic decisions was performed: the case study of one of the most wide ly available construction simulation tool, Stroboscope, was used. Several of its characteristics were found to be flexible and representative of many aspe cts of the construction industry such as the “characterized resources” which allow for uni quely identifiable resources with properties that vary with the simulation time and re present the state of the resources. However, many deficiencies in this simulation t ool/language also exist when it comes to representing the dynamic customized manufactured co nstruction decision making process. The difficulties encountered during the modeling of the bas ic process proposed in Chapter 5 are representative of the barriers that the language c ontains in representing the dynamic decision making process. Such difficulties are the impossibility of ordering the resources in a queue after the resources have been

PAGE 52

52 introduced into the queue, the incapacity to define releasing strategies that us e the statistics of resources in queues and finally the language inability to acces s resources’ statistics in activities unless the activities are cursored. Conclusions and Recommendations The findings and results of this study concluded that the available construction simulation tools were not able to represent the dynamic decision making process e xisting in the real life customized construction process. Amelioration of the existing construction simulation tools or new approaches to the dynamic decision making modeling are required to benefit the industry and add to its competitiveness especially with t he hit that the housing industry has recently taken and the need for more affordable and good quali ty housing. This research sets some basic guidelines for the logic that a constructi on simulation language would be expected to model in order to benefit from simulating the process es and gaining directives for better more productive management strategies. T hese guidelines remain shy of representing a global solution for the obstacles fac ing the construction manufacturing industry. Simulation in many other manufacturing industrie s has shown its efficiency even in a customized environment such as car manufactur ing: allowing for customization did not put at risk the process productivity and did not affect negatively the cost, time or quality of the products. Another main manufacturing industr y having more comparable aspects to the construction industry is the computer manufacturing: various components are integrated in all products in addition to product specific components such as motherboards that vary from one product to another. This industry relies tremendously on automation and uses various aspects of the automation

PAGE 53

53 gains. Future research in manufactured housing can analyze the simulation conc epts used to solve the complications of modeling this dynamic process. It remains that the construction industry in general is a very complex process tha t includes many variances and variables and extends upstream and downstream the production line to include the Client (customers) and the Suppliers. Such variances work against the complete automation and fine tuning of the manufacturing system. In addition, customization in the construction of homes creates many variations on differ ent levels that add to the disturbance of the industrialization of the construction process. A s a conclusion, manufactured construction remains very unique in the complex balance required to keep it on the competitive edge with customization without removing the benefits of enhanced production with better quality, cost and time provided by industrialization of process.

PAGE 54

54 APPENDIX REDUCED PROCESS MODEL: STROBOSCOPE’S LANGUAGE /Control Statements; initialize resources content INIT Orders 5 ORDERA; INIT Orders 2 ORDERB; /INIT MatrlStrg 100; /Ordering the Queue DISCIPLINE Orders TYPE; FILTER STARTORDER ORDER SimTime-TimeIn>2; /Define Activities COMBI Activate; /PRIORITY CutFlrFRm 'Orders.CurrCount>OrdersB.CurrC ount+1 ? 10 : 0'; NORMAL Process; /Define Links LINK L1 Orders Activate; DRAWWHERE L1 Activate.ORDER.TYPE.SumVal==Process.OR DER.TYPE.SumVal; Process.ORDER.TYPE.SumVal==ORDERS.Type LINK L2 Activate Process ORDER; LINK L3 Process End; DURATION Activate 2; DURATION Process Normal[3,0.8]; /Run Simulation SIMULATE; /Run Simulation Until SIMULATEUNTIL 'SimTime==200 | End.TotCount==7'; /Statistical Report REPORT; /Display additional Information DISPLAY "Productivity is End.TotCount/SimTime h ouses per day"; /ASSIGN COLLECT PRINT

PAGE 55

55 LIST OF REFERENCES Abdelhamid, T. (2004). “Lean production paradigms in the housing industry.” Proceedings, NSF Housing Research Agenda Workshop, 12-14 Feb. 2004 Orlando, FL. Vol. 2, 72–81. AbuHammad, A., Hastak, M., and Syal, M. (2004). “Comparative Study of Manufactured Housing Production Systems.” Journal of Architectural Engineering, 10(4), 136– 142. AbouRizk, S., and Hajjar, D., (1998). “A Framework for Applying Simulation in Construction.” Canadian Journal of Civil Engineering, 25(3), 604–617. Archer, A. A. (2000). “A framework to integrate and analyze industry-wide informat ion for on-farm decision making in dairy cattle breeding.” Ph.D. dissertation McGill University, Montreal, Canada. Banerjee, D., Syal, M., and Hastak, M. (2006). “Material Flow-Based Facility Layout Analysis of a Manufactured Housing Production Plant.” Journal of Architectural Engineering 12(4), 196–206. Barriga, Edgar M. (2003). “Manufactured housing industry: material flow and management.” Master thesis report Purdue University. West Lafayette, IN. Barriga, E. M., Jeong, J., Hastak, M., and Syal, M. (2005). “Material Control System for the Manufactured Housing Industry.” Journal of Management in Engineering, 21(2), 91–98. Bashford, H. H. (2004). “The on-site housing factory: quantifying its characteri stics.” (Jan. 1, 2008). Center, B., and Verma, B. (1998). “Fuzzy Logic for Biological and Agricultural Systems.” Artificial Intelligence Review, 12(1), 213–225. Chitla, V. K., and Abdelhamid, T. S. (2003). “Comparing process improvement initiatives based on percent plan complete and labor utilization factors”. Proceedings, 11th Annual Conference for Lean Construction, 22-24 July 2003 Blacksburg, Virginia, 118–131. CityMyors Society (2003). “Affordable housing crisis casts a shadow over the American dream.” (Feb. 6, 2007). Gan, B., Liu, L., and Jain, S. (2000). “Distributed supply chain simulation across enterprise boundaries.” Proceedings, 2000 Winter Simulation Conference 10–13 Dec. 2000 Orlando, FL.

PAGE 56

56 Hajjar, D., AbouRizk, S., and Mather, K. (1998). “Integrating neural networks with special purpose simulation.” Proceedings, 1998 Winter Simulation Conference 13– 16 Dec. 1998, Washington D.C. Hajjar, D., and AbouRizk, S. (1999). “SIMPHONY: An environment for building special purpose construction simulation tools.” Proceedings, 1999 Winter Simulation Conference, 5–8 Dec. 1999 Phoenix, AZ. Halpin, D.W., and L.S. Riggs. (1992). Planning and analysis of construction operations John Wiley & Sons, New York, NY. Hastak, M., and Syal, M. (2004). "Building process optimization with supply chain management in the manufactured housing industry." Proceedings, NSF-PATH Housing Research Agenda Workshop NSF and U.S. Department of HUD Washington, D.C. Hastak, M., Jeong, J., and Syal, M. (2006). “Supply Chain Analysis and Modeling for the Manufactured Housing Industry.” Journal of Urban Planning and Development, 132(1), 1–9. Jeong, J., Hastak, M., and Syal, M. (2006). “Supply Chain Simulation Modeling for the Manufactured Housing Industry.” Journal of Urban Planning and Development, 132(4), 217–225. Jeong, J., Barriga, E. M., Hastak, M., and Syal, M. (2005). “Material Requirements Planning for a Manufactured Housing Facility.” Journal of Architectural Engineering, 11(3), 91–98. Koskela, L. (1992). “Application of the new production philosophy to construction”. Technical Report # 72 Center for Integrated Facility Engineering, Department of Civil Engineering, Stanford University, CA. Koskela, L. and Howell, G., (2002). “The underlying theory of project management is obsolete.” Proceedings of the PMI Research Conference Seattle, WA. Manufactured Housing Industry (MHI). (2008). “Quick facts 2007”. (Jan. 3, 2008). Martinez, J. C. (1996). "STROBOSCOPE: state and resource based simulation of construction processes." Ph.D. dissertation University of Michigan, Ann Arbor, MI. Martinez, J. (2001). “EzStrobe-General purpose simulation system based on activi ty cycle diagram.” Proceedings, 2001 Winter Simulation Conference, 9–12 Dec. 2001 Arlington, VA.

PAGE 57

57 Mehrotra, N., Syal, M., and Hastak, M. (2005). “Manufactured Housing Production Layout Design.” Journal of Architectural Engineering 11(1), 25–34. Mullens, M. (2004). “Production flow and shop floor control.” (Jan. 1, 2008). Narasimhan, R., and Jayaram, J. (1998). “Causal Linkages in Supply Chain Management: An Exploratory Study of North American Manufacturing Firms.” Decision Sciences, 29(3), 579–605. Pietersma, D., Lacroix, R., and Wade, K. M. (1998). “A Framework for the Development of Computerized Management and Control Systems for Use in Dairy Farming.” Journal of Dairy Science 81(11), 2962–2972. Senghore, O., Hastak, M., Abdelhamid, T., Abuhammad, A., and Syal, M. (2004). “Production Process for Manufactured Housing.” Journal of Construction Engineering and Management, 130(5), 708–718. Swaminathan, J., Smith, S., and Sadeh, N. (1998). “Modeling Supply Chain Dynamics: A Multiagent Approach.” Decision Sciences, 29(3), 607–632. Thomas, D., and Griffin, P. (1996). “Coordinated Supply Chain Management.” European Journal of Operational Research, 94(1), 1–15. Viswanadham, N. (2000). “Supply chain engineering and automation.” Proceedings, 2000 IEEE International Conference on Robotics and Automation, April 2000 San Fransisco, CA. Wiezel, A. (2004). “Skill-driven optimization of construction operations.” (January 1, 2008). Willenbrock, J. (1998). Residential design and construction Prentice Hall, New Jersey. Willenbrock, J. H. (2004). “The Leadership/Management Growth Model: Dynamic Framework for Understanding Construction Management and Production in the Housing Industry.” < www.pathnet.org/si.asp?id=1068> (January 1, 2008)

PAGE 58

58 BIOGRAPHICAL SKETCH Diala H. Dandach was born in Beirut, Lebanon. In the year 2000, she graduated from the Lebanese University with a Civil Engineering diploma with emphasi s on hydraulics. In 2003, she earned a DEA, equivalent of a Master of Science, in wat er and environment from ESIB-USJ in Lebanon (cole Suprieure des Ingnieurs de Beyrout h, Universit Saint Joseph). She started working with a geotechnical engineering c ontractor in Lebanon in 2002, where she spent 4 years working with a variety of responsibilitie s: design, preparing proposals and bids, scheduling, budgeting and preparing contracts. S he also had the chance to work for a year on several projects that were starting i n Dubai, UAE where the company undertook new contracts. She then went on to pursue her education and was accepted in June 2006 at the Rinker School of Building Construction at the University of Florida where she was awarded the Master of Science in B uilding Construction in May 2008.


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20101204_AAAABD INGEST_TIME 2010-12-04T12:57:32Z PACKAGE UFE0021914_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 23511 DFID F20101204_AAAXNO ORIGIN DEPOSITOR PATH dandach_h_Page_52.QC.jpg GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
83f9763eb61f05d9969ca2b04c55eb9d
SHA-1
caa350a633d1b08ee96688a1c3405c9615511666
950 F20101204_AAAXIR dandach_h_Page_22.txt
ea8ed4435cfb941f0f5c8b566d4d1138
6e59e15a81d4ed430b434a2d715164be572d1699
102733 F20101204_AAAXDU dandach_h_Page_57.jp2
765be35bcc2184bffaaf5d6a0a53e736
2e119352b8487061fa092e73dcd6a7314e595a7d
6243 F20101204_AAAXNP dandach_h_Page_52thm.jpg
063e52566a299c4326d1d04b739e0beb
d36e74c174c182e3c9636b58f078fa60996ddf4f
1505 F20101204_AAAXIS dandach_h_Page_23.txt
8fc9c379a4bb6b26fc775626017b5939
c2a6b8cb201544a5dcf32ee5be057b6b064a4baf
13196 F20101204_AAAXNQ dandach_h_Page_53.QC.jpg
026d836318be67d4b7bb55ea78144fb7
54545723e532527bedbb77fffacd4d5ddeea24f7
1874 F20101204_AAAXIT dandach_h_Page_24.txt
22d46d113f4003c05849c29838111c0f
c66688a735f07cd379d288e48aa60a5b8e3964fc
59608 F20101204_AAAXDV dandach_h_Page_58.jp2
9e197eb1506ded2bb633d071d47e4949
3757b2fe5ca78f2efe2997a12dfcad10a502adf3
3764 F20101204_AAAXNR dandach_h_Page_53thm.jpg
35148c368911ac2b1af14cea370288ef
cb9b22c78c0e3d239f679e5df7882db483890bca
2103 F20101204_AAAXIU dandach_h_Page_25.txt
f9b25554f78f9d9145cae88eaae2b34a
336270562dece2659ca1c3e270bf08e1d8ed50ce
1053954 F20101204_AAAXDW dandach_h_Page_01.tif
8a51007d6eef07c264894f9aaec2c56c
44673d9816950211df7d2c9ca761742887aef6a4
12973 F20101204_AAAXNS dandach_h_Page_54.QC.jpg
0d6acd7e852a4699c9d0764df89dbaaa
6844f9ab901fabbe2f0af8bd0aeb4b7e0958967b
2526 F20101204_AAAXIV dandach_h_Page_27.txt
d58eedf684339d3430ec4f4e807179e2
d55c50581f991055f7260947641b6cbfc9c89b26
F20101204_AAAXDX dandach_h_Page_02.tif
1c705123c5f736eb8ead8fc05027c146
efd91c14b65a93aeead3a01277cffe655cdd9913
23102 F20101204_AAAXNT dandach_h_Page_55.QC.jpg
f50544cfb6fe3175041725d63d021ee3
b03e01cea77276f6e0fae49711d1bfb00e71f327
794 F20101204_AAAXIW dandach_h_Page_28.txt
a510561e0c1f7c748dc43cd048190778
e29f09e9d7ff51a211fe70e0dcac9ec4c3095be6
110021 F20101204_AAAWZA dandach_h_Page_29.jp2
1ce32ec7b69fdbb3fd351ce9ab619c81
586290970c005e31320585d746f494e6a0483342
63867 F20101204_AAAXBA dandach_h_Page_37.jpg
266e041f2adcc6e5aba392892d153257
9830df551d60af9fe8417b63a2857d83498426ec
F20101204_AAAXDY dandach_h_Page_03.tif
5a6fac1c639fd18c5f7654fec9ac5f3e
6024f0215917469dc933c39416447e232e7e278f
6327 F20101204_AAAXNU dandach_h_Page_55thm.jpg
01c0077491a7c6a21f0d4fd544d943bf
e6790f9504157fc2b113c37d3c2e05096551dfea
2055 F20101204_AAAXIX dandach_h_Page_29.txt
484300315597d71677247f818909aa32
3acca35954dd165b0ef1a07b70b024c339bff391
5798 F20101204_AAAWZB dandach_h_Page_02.jp2
16699f5ff1920dc2f7ba25a1ea5dd646
fa18bb3a6b68e8212217f33326a7c4d0139de380
76300 F20101204_AAAXBB dandach_h_Page_38.jpg
5efec97dc5b487fcccd476bf96f82c4e
bf07bc0cf78da26c828b9a17abd13218d86d76d1
F20101204_AAAXDZ dandach_h_Page_04.tif
26820d65e842cf867a524f00a5ac0d1a
242419e89b0886b529973c89a1da18ad37d7c196
23453 F20101204_AAAXNV dandach_h_Page_56.QC.jpg
cd0ad2260a18da9db8770ba369b4ffbb
ce8938a26201ea69e5339f6b2e035fb5761a2bc2
1957 F20101204_AAAXIY dandach_h_Page_30.txt
b661821ce595bd67f65ffd876f27f047
852ae3b1e5234b258ad871634223e72a0bbdf95e
59899 F20101204_AAAWZC dandach_h_Page_41.jp2
692daa7f83378ebd8add071a0618b2c1
83ab28cdd9a3d583066096e8a75670ddc05baadb
70055 F20101204_AAAXBC dandach_h_Page_39.jpg
d1fa0cb64b5ab07561ed3dbe014de6e6
595dd0406ec39fa4c95e78e390a94050d672ce43
6228 F20101204_AAAXNW dandach_h_Page_56thm.jpg
c269e69dc2f349281edd95de132c8920
c7a3a3d5c1f1b64ddd15a2195e8086826afff18f
89239 F20101204_AAAXGA dandach_h_Page_04.pro
bc0cf9e9b380b8ed6e26f33daadb0f42
1dedfbd9391585b8846a0e1755680ccbdd81afd2
1864 F20101204_AAAXIZ dandach_h_Page_31.txt
4f8c1de7337e1585e1238b474ebc0ccf
2f2ddd6803686caef76c964e787991d763256bba
158 F20101204_AAAWZD dandach_h_Page_03.txt
1e376b5b991100428fd45e24e0a43715
5d163f105fb289f8efa247f8ba6a34e65f48f8b4
70986 F20101204_AAAXBD dandach_h_Page_40.jpg
0e1d485f2f33a2008ecbc31a1d4e8137
9811339aab332d340d044790ebf34dd60fdfd0e2
19931 F20101204_AAAXNX dandach_h_Page_57.QC.jpg
c5abacdf86608a2861944bd48a12a4d0
196e33a9a1a108edd4049eb6c35d1b2c95597582
28060 F20101204_AAAXGB dandach_h_Page_05.pro
0957f02bef3cd8f5c1e060895a054ce3
e0bc1e6d6648d465ee451c31127fe11a3f9a90b6
54273 F20101204_AAAWZE dandach_h_Page_44.pro
c47591c9fb6dfb2e3efcf64b68e118a4
35551dbe527d86bc2e8b2143596b75a56e94460c
42622 F20101204_AAAXBE dandach_h_Page_41.jpg
bcdd48b56faba34b50a2374f1c3c1453
ea0e7ae86827b8bed22eb915cd3b9b0c4d7bed98
5570 F20101204_AAAXNY dandach_h_Page_57thm.jpg
26e28eb372ebd8898f78ce572d6ede23
74254271f04b7e79eb38a6060aefc7316ef4de2f
19850 F20101204_AAAXGC dandach_h_Page_06.pro
1b7c0c50f39886846e3558542ce82257
9a5e1f5c41c5a4ffc470d64f59d1e63b6f431168
883 F20101204_AAAWZF dandach_h_Page_12.txt
c91c0107c077041aed0e70dc8b458782
e8135f89cfdeff109ed6e9bf1a0b2cca8a1ab1a9
75074 F20101204_AAAXBF dandach_h_Page_42.jpg
c86c6be40007f023a704a5a4aca90aa0
3d1abf64d26e8fb2d0c37120b5eddec8e299e8c6
13831 F20101204_AAAXNZ dandach_h_Page_58.QC.jpg
9cd899bce0786adf2914df12f1999800
a542a721892f47206a6ce3489b3c351205b6dca3
44569 F20101204_AAAXGD dandach_h_Page_07.pro
e1ae2f19bc76ad1f34192767240cbb7a
8b386be8c82e1f950f15290d7f2b4d4f26065d9d
54024 F20101204_AAAWZG dandach_h_Page_32.jpg
28b9fad72aa08845b6eda0b03e8ff510
f3c8161d5580a6be14e1b8a8a4f693b4327bf3eb
72739 F20101204_AAAXBG dandach_h_Page_43.jpg
5909b09055d7e5ccc8616f3cada36d82
ae59f9a55fe1fbef6af383e68a0ead7ecd098679
23055 F20101204_AAAXLA dandach_h_Page_17.QC.jpg
4d17954f8791da3949048fec604f4e6c
f70ad62e72d91b9c814ed4a1abd5552e2fe4dcb8
3707 F20101204_AAAXGE dandach_h_Page_08.pro
b31e01ca9ec4223494933cc4adc936f9
b7e9953cf3e5bb26231acdba08b14ba356443fcc
5897 F20101204_AAAWZH dandach_h_Page_34thm.jpg
6d02069ae06be4f20c7d8bb237f62b3f
da68560a66acd0b620e1a0450baaa8d9dca64430
74917 F20101204_AAAXBH dandach_h_Page_44.jpg
a76165a0561ac3ccb41089caddb36d04
cd23d6509c8e0c40c1a794dfcb00a06088be0ab0
6298 F20101204_AAAXLB dandach_h_Page_17thm.jpg
23f6474ef4a722b1c340e4a2cc6e6c42
9d2f0a76fda2c5243f416f8eeed53e4eebad988e
49425 F20101204_AAAXGF dandach_h_Page_09.pro
cac974b5a413c27bb10a47868db6fe0b
14ad7f38e32e161285f6e8f95ed2eaebd3312f56
53735 F20101204_AAAWZI dandach_h_Page_22.jp2
358c0f31fa0726ee3a62916516ad696f
c5e2de1b9a758ebbdd5bcea1b8a8c00e8bcd2412
67845 F20101204_AAAXBI dandach_h_Page_45.jpg
6f8951c24e878426947941de7c93da2c
fe05fcf6da35a0e8704a3d97412dff0603b63717
24965 F20101204_AAAXLC dandach_h_Page_18.QC.jpg
b57ea7e81ec62494114d031dd805ea53
71cb9ad544475c635a3125d70ad7d86130dda6a9
47218 F20101204_AAAXGG dandach_h_Page_10.pro
8c6e96dc72d912bc90574c4d1100c5f3
471044593cef96fb0e409041a39916ec71fac977
2125 F20101204_AAAWZJ dandach_h_Page_26.txt
df748a0ae0d6e776041ccb910a817818
d9b3d539c4cbf6b55fd306d81d1bae4db24a77ec
70174 F20101204_AAAXBJ dandach_h_Page_46.jpg
aedc2f448e524bff0932c8a8127cf784
02dfcc72e8ab906162785081c2c1bdb8de7ebf3c
6759 F20101204_AAAXLD dandach_h_Page_18thm.jpg
7455f1a0798cca9062405c8fe047a168
26115564d6e901db5cd5fc112869149a88af01c8
52046 F20101204_AAAXGH dandach_h_Page_11.pro
fdb3fa740dd687a27a25ac66da24ca43
3c5aa6d31cb5281bb6868adbb3f282be37682d8b
6509 F20101204_AAAWZK dandach_h_Page_13thm.jpg
dfbea802234b6a464d550f0209aa7860
36d67b4b781c86cc02ebd5ec08542ae5ab6067fd
69011 F20101204_AAAXBK dandach_h_Page_47.jpg
c4f5f60f3456156c8ed73c21755a3171
bb6c324af892e30e00f8c370da841bd67c600cd1
23811 F20101204_AAAXLE dandach_h_Page_19.QC.jpg
a5157afd98317e598adbb149c9bc656c
3aa564a368ed1aa7b4c17fefc1523c553b453c23
21820 F20101204_AAAXGI dandach_h_Page_12.pro
c2ec94983f0c9519c20351d4f0edfb29
05f03d508ec75405dbd22e7570b3ab0ab9abb8d3
F20101204_AAAWZL dandach_h_Page_39.tif
0cde9b410913db9025f289f03cbc687c
6f79ee6bca3c9093e425f957dd24c3b6e219a920
62200 F20101204_AAAXBL dandach_h_Page_48.jpg
1a0a9ba14010bc6732972e230e4b3828
9d6391b794d11f8068eb28f32693ffe13f734a8e
6536 F20101204_AAAXLF dandach_h_Page_19thm.jpg
9d5e76e5cb65ab6f4b6c60bac8d85248
47d98bad8dff43cbcbf68eafea7a30c3f1bba222
48581 F20101204_AAAXGJ dandach_h_Page_13.pro
e455fd8fefcf8727288b91ceba05c231
640393b4367d6395091ab5a5a0386730294604cf
39531 F20101204_AAAXBM dandach_h_Page_49.jpg
841943f3a0eb41c041ac3119b70c32c9
b3811a9ab6320e1ffaf213df8776d37df5c24a33
23377 F20101204_AAAXLG dandach_h_Page_20.QC.jpg
1072c43b81db2fa98ad39c604b8f80c4
ea67698181c37269d0be2d5cfdae6fc5df411c04
50856 F20101204_AAAXGK dandach_h_Page_14.pro
33ec4cca15fba3ac3951335cdd6c92ff
a36f5e1db16a82a7f76ca5c8eb853480be13ae92
102638 F20101204_AAAWZM dandach_h_Page_10.jp2
c10db581965cac14e69c86182cfc5f41
78b8246e3601beab3967fe0788523f117c17d578
68022 F20101204_AAAXBN dandach_h_Page_50.jpg
43d8e32fc667da4d05119cb61386e6c0
e4cd91b5b0843d511dad78951b0ca26c1784aee5
6392 F20101204_AAAXLH dandach_h_Page_20thm.jpg
49ebfa081d42459f7f1b1f8c158b5c92
e8b5d9295dc0057af60d481e125959a0dc7c990d
48744 F20101204_AAAXGL dandach_h_Page_15.pro
0a4a0dba0f114e6f92c0c33601a26e6a
4d30297c05a1de277c069436ba9d2b9329e5d027
25981 F20101204_AAAWZN dandach_h_Page_06.jpg
767d885db7a593cc8f85a7e9325acd64
6e91feb9cda5a2f3cd50257549f0a389b615693a
73466 F20101204_AAAXBO dandach_h_Page_51.jpg
fc396aa214ef5b6ee1cabea49c9a75d2
9666f25acd715b7033bcf8e2ef3c3cb00ee06eb6
23604 F20101204_AAAXLI dandach_h_Page_21.QC.jpg
1e875bc307d530836a497f9a004c1441
d46b6c434dc6936eed83f391ca875287b5bf544d
49139 F20101204_AAAXGM dandach_h_Page_16.pro
2c9e449c7e0e8830cc0d3a58341f7005
a8c3516ceee4dddeb5e8b57def72b368d9007be0
22881 F20101204_AAAWZO dandach_h_Page_24.pro
7b20541836485c57b14745578e492dde
1df258264a3c93afad5c9d1901ced753774b46f2
70307 F20101204_AAAXBP dandach_h_Page_52.jpg
cdce08b0102fe73288c0112dfd0a8c8e
76aa78e41c5247315be84d0e934fd5463ea78c46
6477 F20101204_AAAXLJ dandach_h_Page_21thm.jpg
2fc76856788f57ae04ace9cbe5c5ec62
6ec2fd66a8984a18c50e9d0ec02c5db66d564701
F20101204_AAAWZP dandach_h_Page_49.tif
b25d1c12fc9e9c2b6ed4ddff7b65e39e
bb23bbb6b5534268852af5260dc6063666058357
39881 F20101204_AAAXBQ dandach_h_Page_53.jpg
765d634c6b18f6f6d232020949cdb6a9
29326edbe3edf46b3ddecdc4becdb8dbe8c4db37
12798 F20101204_AAAXLK dandach_h_Page_22.QC.jpg
103b84299b888f5d5aaf057a193254fe
e82a8b637a8b5bed9998fd7cc35331d9b26bcdc1
48734 F20101204_AAAXGN dandach_h_Page_17.pro
87e15861cfadc0be288e8da1555ff199
96ce49275ce92597ad177eb9a325bd8688f0c134
F20101204_AAAWZQ dandach_h_Page_53.tif
f5705674a05aaceec628451e0cca26f7
99b9940899e6f28189129b322cd7f9e07f64fc43
40676 F20101204_AAAXBR dandach_h_Page_54.jpg
13c087da2d489d43d7daf956ea508842
00a3aa62655bb7b7d3ac35430b14dd7b5bbd6320
3700 F20101204_AAAXLL dandach_h_Page_22thm.jpg
a4a90001a10c3e51b9fda4c66837248c
5f7a2935233ca7b6af7cc29f2354307954190c69
53571 F20101204_AAAXGO dandach_h_Page_18.pro
ce426df2fb7f0f3b9820b2f5a3a8b273
6f52d1dfeb323bca23056a4eb566d57f95164205
3382 F20101204_AAAWZR dandach_h_Page_12thm.jpg
96a28a2733a0f5e46fb97f5d75af1a8f
f8e420dbf5e956e676b1dd6730847afb642d24a2
79820 F20101204_AAAXBS dandach_h_Page_55.jpg
ae48c9b75fa298748cb4653556a08113
285056b5b431449c741e5c43f3e5e8226fafddc7
17581 F20101204_AAAXLM dandach_h_Page_23.QC.jpg
8c25d9a59a75af474c0bb01c105f39f1
d77ffe66821fc887587eda8f556f02ac56069a05
50919 F20101204_AAAXGP dandach_h_Page_19.pro
8aa7d6b48ececc6626c7f8da406fd5b4
2b1082ce510599e7efc481411f053bfee4a67c51
89602 F20101204_AAAWZS UFE0021914_00001.xml FULL
7281e7db487f8763a70b66f57847161e
b6b8aea8ac43e008c437302da51ce0e5bc9288fc
5027 F20101204_AAAXLN dandach_h_Page_23thm.jpg
fac738f9fecbeff9b23dd4beca95ea1f
137066a6c7267ac090ee7a086d35ef5b61d7cdb6
48915 F20101204_AAAXGQ dandach_h_Page_20.pro
ffa6aff614249431d63aea6e0acc1ede
c3494b4e076e850dea634219096ee7d5a4a8f8ef
80354 F20101204_AAAXBT dandach_h_Page_56.jpg
3fb573bd31d17ead845866991b306dc5
d1eb6bd76a0fb9c1b46f940242f9563db9f99994
11687 F20101204_AAAXLO dandach_h_Page_24.QC.jpg
a5f9a45e9eb54086b8e13c1493ccf10a
84096fb1144d3bcd43501ee6a6514a073712c5a8
49757 F20101204_AAAXGR dandach_h_Page_21.pro
25b4fb37b84b9150d41b3c2628d0945e
661764e5e0e70300d0ce6c6a2b71a40651d96d54
42946 F20101204_AAAXBU dandach_h_Page_58.jpg
60db91a048a9708efbea157784f1c351
d8f1c3ecc52d6e23b42cef4d7a43b53f148c5a38
3643 F20101204_AAAXLP dandach_h_Page_24thm.jpg
23652a5aecef63dd8f55643517379ac6
09f1118a5c00baba6bba7ab610ae156e50db6215
23733 F20101204_AAAXGS dandach_h_Page_22.pro
d588d643a4bf923d7c47a0eefe76f3fc
e093c6f72f473f749763c9d88f2913ac6fd15de8
23458 F20101204_AAAWZV dandach_h_Page_01.jpg
82bee77d24c8ee0722b12d378e146fc6
6cd0173b92b6dbfb5843bc313cf2d54fdde9e86f
25288 F20101204_AAAXBV dandach_h_Page_01.jp2
95726e5f0d8a1253385d2fb3f9cf3e69
206ec3031eceddb215d376ef40d486c627c2e82e
24651 F20101204_AAAXLQ dandach_h_Page_25.QC.jpg
88ee2e87c72d87031a8e2f53dbddb486
c8a594a3406469b3969ba6759cb50b786eb81bb8
35980 F20101204_AAAXGT dandach_h_Page_23.pro
618dc67de1ec6cf71332962d7dc6194e
303fb7afb17a5ad449ed6431513f2a788a559772
10170 F20101204_AAAWZW dandach_h_Page_02.jpg
2a98e49968373de571562e62cd5cc574
bf0cc47de6a6282e912ebe5711f6b08ff003de8b
9512 F20101204_AAAXBW dandach_h_Page_03.jp2
5aca56c19efb4ea435847444f5e56514
d9f0c36ea78e72a63e96d913fc20f946f55d9e3b
6635 F20101204_AAAXLR dandach_h_Page_25thm.jpg
fcf3b7b6564ee483a15131144dfcd710
56ebd26edcb85d73f743b3a795311f23d6d4ec71
53819 F20101204_AAAXGU dandach_h_Page_25.pro
3d883781a6dba2fd13ac8d6218988e83
3c0c3fd83d4c89071c675658da506d3942c2f07c
12419 F20101204_AAAWZX dandach_h_Page_03.jpg
0303c26bcb42f02ee3a99d3dafdc3e3d
8481d7b3de3ac3341d89254f42ea307b9a782ffa
118061 F20101204_AAAXBX dandach_h_Page_04.jp2
84880ebc0dd48013b54595f90f91bc85
66b2da374b6a2b8f53ef37a2a65b74e5e17f7f92
24459 F20101204_AAAXLS dandach_h_Page_26.QC.jpg
9b6efd7ec2899d40b46b17893d738aeb
6581cf87a9de24c64a7837e62db870a2d6aadcac
53973 F20101204_AAAXGV dandach_h_Page_26.pro
2c39f4bf17eb720ed85707fb7cd6abb5
d32145fab5cf525c46430a9211b888e4e34d83d5
83988 F20101204_AAAWZY dandach_h_Page_04.jpg
336a132e6cd55a75fb95b4ee429c007a
a3471e11a8a1042f2373d2f6e084fc7f150515e0
37104 F20101204_AAAXBY dandach_h_Page_05.jp2
2e58113d32246a03e04e0f2307e6d465
b07f0eada94cb05aa8ef5701dca3641f69fe8238
24950 F20101204_AAAXLT dandach_h_Page_27.QC.jpg
81356dc834dff4255c2a6c308637ec6a
fc40945fe3ad2ff2855d63109df99bf70f43845f
60107 F20101204_AAAXGW dandach_h_Page_27.pro
9022e3049ce0f03cb5c8488091621487
45d3e78728f19e6411fc323ac4ada0e60340fe35
29569 F20101204_AAAWZZ dandach_h_Page_05.jpg
f9b077ca39a915a28b2a0c2f21955b4b
537f5b54ba84c5680852481714cbb8ec98fbb8b0
31108 F20101204_AAAXBZ dandach_h_Page_06.jp2
7f98ec9e29048d6b239c64c44735e388
11a393740b0456d2eebbe0330a8289f1acf0c74d
6680 F20101204_AAAXLU dandach_h_Page_27thm.jpg
4228217e4706029a0b84a0ce33fac67c
a2a9a9a6970c15a3c2dec7f51e9ca16dfc304ada
19984 F20101204_AAAXGX dandach_h_Page_28.pro
1801e14e311f0d9c8cdab88e2bbac8de
5cfa9643fe72ecfb83b7e964cdd309a7d28f619c
10931 F20101204_AAAXLV dandach_h_Page_28.QC.jpg
a8c67026df117acb4ecf96cf6103d284
4ca637ffe42ba3ea70ea89a2d6539f2c19690189
3348 F20101204_AAAXLW dandach_h_Page_28thm.jpg
02e7bc2263fadb6fedfa7d118435519e
4b827caf1c9a553d0fd28879b3b7717bd5a887e0
F20101204_AAAXEA dandach_h_Page_05.tif
031b51a95c92086cff291d3334a6bf01
8c9416b19ce75172be95fc8069326aefe240c98a
50529 F20101204_AAAXGY dandach_h_Page_29.pro
df344c85b3d1366a82c32552b9c1afcd
a891d88f3303e6dca3284c65db5c55b82dbe1a64
6553 F20101204_AAAXLX dandach_h_Page_29thm.jpg
e864bc9291f08303dbd2c77d59e035f1
6101fab5c05e8029ac60add477c7d0e4e0d2888f
F20101204_AAAXEB dandach_h_Page_06.tif
027fd1e580fb1b581779991454a65674
406e464e6ab10db323881df85922d4133bc961f8
47418 F20101204_AAAXGZ dandach_h_Page_30.pro
aff0789e2ba984b35c1d858508e2b20c
163e33545124d7d2a0337ae51f39ac5eaaafed01
22013 F20101204_AAAXLY dandach_h_Page_30.QC.jpg
fe71d62b18c7a2e359411338625125a3
597918c1b0e9a315562032d16f47b22050d0b538
F20101204_AAAXEC dandach_h_Page_07.tif
6e896baea51e35c1fac43f778f525e76
0a6595567e83e02dbc746782870c757fbee4a04f
874 F20101204_AAAXJA dandach_h_Page_32.txt
1eaffb9d4d76e656c94092f5a4088421
62446bc500191a013d8854730c5a4aa1ec2c12df
5944 F20101204_AAAXLZ dandach_h_Page_30thm.jpg
58d61d405a577d596a74b4ff92947498
7b126d0cdf33b4b0680ae2d65b0b56c7ad1d8703
F20101204_AAAXED dandach_h_Page_08.tif
38d9377007e95e492b4e8b50550083fd
3618329af9f197194263f0af4a0deebc6ab22535
1983 F20101204_AAAXJB dandach_h_Page_33.txt
0d0a7ee4fbeb34bc2b2391f245091dd8
aa17979e1af3efa1f2cb8ecfd807be72c1a75cea
F20101204_AAAXEE dandach_h_Page_09.tif
a6b414763d63edc4823668905077e2b5
21a44b753126d1071ae8e25373a0d98feaa9d5cf
1829 F20101204_AAAXJC dandach_h_Page_35.txt
c5fc44058d676fdbd8fef5a77a1b62bf
f38fa8ecd6de386e9969952759e2c2c99b0e6c93
F20101204_AAAXEF dandach_h_Page_10.tif
f5634a94b578ada6bd768275a913b833
28de45798a7af7730f0cbfdeef21ff1e53d0eb11
4002 F20101204_AAAXOA dandach_h_Page_58thm.jpg
b1823a2e9b57b7a40a86fb7e7cdfd922
d3c367871b938de6891e6a02a4d517526399b942
2127 F20101204_AAAXJD dandach_h_Page_36.txt
929e9c417de207e57cfe63d248dd2ae3
7c86f4601656cb418a42c9355716c5aeff1732f2
F20101204_AAAXEG dandach_h_Page_11.tif
68f44957a2f41e51ce960e36735867ff
e3896d583d8474d4a1ae72536c7a03c2f529f945
69362 F20101204_AAAXOB UFE0021914_00001.mets
8faff23f18521b774f315bfb5edbdebc
917a32a41170b73d6e447803208c85d7e02d20d4
1893 F20101204_AAAXJE dandach_h_Page_37.txt
2d213f69c6d658a832b24007ce2b3a8a
901d2db3b4f629e14bfb565e42ac6218771076c8
F20101204_AAAXEH dandach_h_Page_13.tif
1c21fa855ec3b75541e522b200c66caf
626b13d868c93066cd5c7c04a18387c68e7f37e6
2153 F20101204_AAAXJF dandach_h_Page_38.txt
291c3f6e1336eb3e4997233b953b933d
72fd870b4e4116cceab253c4982daa63d4ab7212
F20101204_AAAXEI dandach_h_Page_14.tif
3828fed0dbcf1fffd9464212f8590f03
67a1f330ed95c0452f7c7321e16c57cc6ec87ce0
1978 F20101204_AAAXJG dandach_h_Page_39.txt
5565e07bfe2b2f957f1209f1b43a3cb6
08a8ab8b10326bb093773ccbc1b49d6b1e542a6a
F20101204_AAAXEJ dandach_h_Page_15.tif
12b40a3ab099855026a289555e479012
e5004b386bc893aec4f50472d68050dc8861cec1
2021 F20101204_AAAXJH dandach_h_Page_40.txt
1d3a4dc9f3623d0f6f8a3faaa1350d2f
502567218af86a0b24d2edd7e26b59ee7d7e733c
F20101204_AAAXEK dandach_h_Page_16.tif
a015e46bf193be39bdc9571f9ce469d0
18933a2a5859fbad1e187a9343c66b1a5909d398
1089 F20101204_AAAXJI dandach_h_Page_41.txt
65593efc5b2321b95d0cbe9f249dc9f5
cb08283333f6af2340ba05aba0e9d6c3c7abb322
F20101204_AAAXEL dandach_h_Page_17.tif
e3a1ebe65e3c462549f9d0bda7f62091
fbaea6077f4daef296f7fbca7024e226a66f7a71
F20101204_AAAXEM dandach_h_Page_18.tif
b235dc9892c2dc69aef763464008083a
737494cd601ca030bb8424cf660a20fa2bc989b5
2105 F20101204_AAAXJJ dandach_h_Page_42.txt
5227e8a99c0e087931da571f7ac6f398
cb7d58fc37ce909edc335900b41302af55cd4637
F20101204_AAAXEN dandach_h_Page_19.tif
11d7241986bb4d27ae801db9c0774133
26335433f4fc55f5900254de95d3f598f1b1c3db
2098 F20101204_AAAXJK dandach_h_Page_43.txt
504b6ed9de6acee6159d16e327ba7681
785aff53259c19c3ee4fa5cb691028bf51184132
F20101204_AAAXEO dandach_h_Page_20.tif
9a6b12af23ba9249e2b947a8f6ce3726
3b21a0aee7e30ac01e325d132e12b10642be7dde
1765 F20101204_AAAXJL dandach_h_Page_45.txt
49c38f4ee4d7bb007f7ee24e1875911f
95b337606b11d4421372753c829cb3d4acebeba9
F20101204_AAAXEP dandach_h_Page_21.tif
dcd2a3b341e9f371d0c209211ef18643
90e3ad7c2e2f08dda2c29dff3745d221e9ad3b9a
1984 F20101204_AAAXJM dandach_h_Page_46.txt
6ebbba7e38be11595526d95d2328b773
3f20b4790d3b4bce048f8f7d5089761a58baf6cf
F20101204_AAAXEQ dandach_h_Page_22.tif
ff7de25077a5dc2f6cbd1fab9e4be0e6
5336572c165aded7aaad1263dfa04a83600bc55b
1942 F20101204_AAAXJN dandach_h_Page_47.txt
1a5764a8a29c82f4fd66a2527d055f1b
39dd6e2c90f8711da7da3c66ae8ef6652d0cbb93
F20101204_AAAXER dandach_h_Page_23.tif
d8e2529346dbc5219bb4b828ad360f00
61e1ad4f22d1940fa6b29ca665afeffd0413dd09
985 F20101204_AAAXJO dandach_h_Page_49.txt
92a1393f635ef0c02f23570f20e604e2
c1cae9d7ff3405eefad6c59d8741a526250baf5a
25271604 F20101204_AAAXES dandach_h_Page_24.tif
2afabfd56503ff879aa2eaf000d251dc
af0399cf516b672464cd6f13299f627d23623052
1989 F20101204_AAAXJP dandach_h_Page_50.txt
8aa4a1eca1ff6f3aebb0d8a6532bba8b
42d10889812c45e35d02f0f6d497443a10513f40
F20101204_AAAXET dandach_h_Page_25.tif
7c02433896aa1918fa5195be3fb66c81
3b3765032517360f1663dde16d7dc3456263fc4d
2036 F20101204_AAAXJQ dandach_h_Page_51.txt
d253c46b4c6c7071f6307a797a1fa4f8
edb964ea3714746b61f7c5151991f30ff704ae07
F20101204_AAAXEU dandach_h_Page_26.tif
7be944bfd61e02bca3956f84481976bd
17b7a3bc3176f53a0f51b7883f441e3f8614c2dc
1945 F20101204_AAAXJR dandach_h_Page_52.txt
9462d75fb4192e3c5fb0206b7cedf845
3b405a506b382422412f6a3a468dec5aa05c6338
F20101204_AAAXEV dandach_h_Page_27.tif
27eac8fa0eebe6d82965ba63cc0180fb
228b1845258cb1c74ffb2eb600708118e05981ce
993 F20101204_AAAXJS dandach_h_Page_53.txt
f0026b962880b710a5e52d78d564a6b6
67b04ae10e53d655946443db07eebb90ed602acf
982 F20101204_AAAXJT dandach_h_Page_54.txt
db3e6b57e2e3dd10446a97554b271d8b
c032ac6dcc6f6b3cdd55c2cde7d378dd9bdf4772
F20101204_AAAXEW dandach_h_Page_28.tif
f62499f7682c650ed9881d44ef7faab7
c5df7eb175f13e2aac8f3de302f1a427d70bea52
2262 F20101204_AAAXJU dandach_h_Page_55.txt
81ebea36be4274c40a3c24db2daae9ba
e9edc1487be644840ebf6c1d41bf697d46901439
F20101204_AAAXEX dandach_h_Page_29.tif
2bb7d3014c6a0389c38737a52a57f629
81a4dfd13fc819e9607701917956e421ad9d67fc
1846 F20101204_AAAXJV dandach_h_Page_57.txt
0060682d155483de6fd7a4f7cd21f241
80d958e03a7156c710f23987a6bbb0b320b423c3
F20101204_AAAXEY dandach_h_Page_30.tif
aea5d505b9962d2db39e85ffbddf6960
6e84636f6966040662634c439f3d4583bfeab9ff
1062 F20101204_AAAXJW dandach_h_Page_58.txt
04d1b48addf040a1efa14e511063a620
1ef733de71f1cf2a702d9fae69adf0bb8fe964f0
96524 F20101204_AAAXCA dandach_h_Page_07.jp2
ec7f630cb8533682868a121b598fc3ad
1155be4d99d9a381850318234efcfdf081a3637a
F20101204_AAAXEZ dandach_h_Page_31.tif
976ed842e2bd4f006071b8297d905b25
dedf0cd7f91d80c6770b4439333ad8e22b598d45
219237 F20101204_AAAXJX dandach_h.pdf
f1e943239167643a8c1c57dbc05b9110
7c4892bd032af768db5ff5be18dd8cd0bba00662
11267 F20101204_AAAXCB dandach_h_Page_08.jp2
4547001105b9c85d18d726e4027bac72
66dddf7263a525e9b25e897fa737895161098cbe
7081 F20101204_AAAXJY dandach_h_Page_01.QC.jpg
82bacd3793e38c021e9109a995c4c4e7
ce6f08cd3c7ead0502f87ad95c81dbd563e97293
105618 F20101204_AAAXCC dandach_h_Page_09.jp2
99f142a7ee8dbaa9d503c3bc4d6cdbed
99246276b94b77cc824099fa65585c1a2b4d8817
39759 F20101204_AAAXHA dandach_h_Page_31.pro
560fddf84f150a3635a8833cc2a62cc6
6a28292bab809350e8bf8657e0d2a3a887815525
2457 F20101204_AAAXJZ dandach_h_Page_01thm.jpg
c08feae8992e40c39b9648d58b885445
1a575d727e8b09e2d4d84609a180e7433baf28cd
109828 F20101204_AAAXCD dandach_h_Page_11.jp2
b9c0c7118d209a4c101e6e6e94f0415e
3cf77675238f78ae441d32ddd6bcb6541e3b6709
16944 F20101204_AAAXHB dandach_h_Page_32.pro
b6697597353eab2766756bb56bacc717
67efff8b524c77319cd05ff0d01c2c0a0f1ee60e
48132 F20101204_AAAXCE dandach_h_Page_12.jp2
41dcb50e116397d64597b4567275452f
ccfd5be638611b0686dc8ce24f7c02e32325fd7a
49730 F20101204_AAAXHC dandach_h_Page_33.pro
a77bb7f087e7de48a8da44f68788ba3a
75a3f0c4978e1446eb1dc95d0e1797e8c070ed81
103920 F20101204_AAAXCF dandach_h_Page_13.jp2
59d1a9097732212dec5c8c48660a476d
6cfde334d48c028f8f27f0a54b641bde68772dc8
21496 F20101204_AAAXMA dandach_h_Page_31.QC.jpg
3a2977faa51d075410426574d1dad1e7
d8adf0da437272953510664f0ae6f3d507d9734f
44614 F20101204_AAAXHD dandach_h_Page_34.pro
ac08c6dd75bab353f9bf3618d7622631
58256d80f3a114de26b413cfbece8a4354f139ea
107433 F20101204_AAAXCG dandach_h_Page_14.jp2
405848d3310c2816413c790e86e10874
55ade7fbb8defb4f2596e1557ff0c7fb8180638f
6128 F20101204_AAAXMB dandach_h_Page_31thm.jpg
4c14e78e1baa553312ff903e9eabe2b2
d5267d8566b80473e282d7d3b9fc05a49f62eff7
44738 F20101204_AAAXHE dandach_h_Page_35.pro
8168675d54dfd14c8c04a6345a12ecbb
409191620032be151874b5777a8fc94cc28e862e
105962 F20101204_AAAXCH dandach_h_Page_15.jp2
cb3f75cadddd5e49be78d418cc6efb5e
e6c528135b7f3736f780cff757624c4534506da3
16462 F20101204_AAAXMC dandach_h_Page_32.QC.jpg
e7d7ca64016c42338063357fee0315a6
db2baef1e9989707a3470f7b88a362759812da89
54298 F20101204_AAAXHF dandach_h_Page_36.pro
3d1af4e5bf361e5f1b0c11b1d991223f
4ba3b973be4da7178bae970669c9313d00debe9e
106791 F20101204_AAAXCI dandach_h_Page_16.jp2
6f1340e72981b214ded21d5641c68eb9
5c4c345dd3b1862a1367eec63c450505bbd056ad
5265 F20101204_AAAXMD dandach_h_Page_32thm.jpg
32f2d4614d5eb91c5c895ffc2510d9bb
47cf3d857a96e23b1ebe32d7eecd8e158de82678
45287 F20101204_AAAXHG dandach_h_Page_37.pro
282ed122e0e3d26baa8e70557c9d370b
78971f15ff8f99b5f684546f7d21659f02070469
104461 F20101204_AAAXCJ dandach_h_Page_17.jp2
2896e0f4bb2fe44fb6c9f8a6b5bbb786
baaf85055b97c03782dce92950f944dc7beee831
23070 F20101204_AAAXME dandach_h_Page_33.QC.jpg
b146ea1eb38cf87aa2503ac87f59cdc5
fce80e7d6e952246b30b34adbad4f18ef814a32d
54857 F20101204_AAAXHH dandach_h_Page_38.pro
4f5e050c9cf7940bf67f1682d9d651be
d19c11747d848957aac00a8b9f4ba6e82091de48
112227 F20101204_AAAXCK dandach_h_Page_18.jp2
349b0eff7f4ec34b0650ecea02dfa2f0
04c486c648bb3ffc0160c4f89e8fee1911e9e577
50313 F20101204_AAAXHI dandach_h_Page_39.pro
1654ab95142a2ffa55e6b2d72dbb52a8
ca2d06150fc6b7806eec6774237ac05d7ec6877d
108878 F20101204_AAAXCL dandach_h_Page_19.jp2
e11b17dba6a244e9e4f0803029e1ba48
280346c19bec06e10b8b6a24dd151e0c3dbb4bf7
6433 F20101204_AAAXMF dandach_h_Page_33thm.jpg
a1329ea9cdddc9a847d6390a1066b23f
0c4eda4523b00d4fa2fed70ea6aca7e55c4a3c47
27247 F20101204_AAAXHJ dandach_h_Page_41.pro
5dd117e5dd20f7e0e7d7500c0599df45
2439af5aea7836eaac114a8927a1c74900e7089a
106020 F20101204_AAAXCM dandach_h_Page_20.jp2
6ce85c85be0e015dc0899689d52ac2df
8ac3f9d76ba9f78a65c00e8c85025cc76d15d015
20981 F20101204_AAAXMG dandach_h_Page_34.QC.jpg
e3c06e1c11467be76d4e93e329ce0c54
d9be2cc69cfebd8fcfa1ffad68ede19d9afea9c6
51757 F20101204_AAAXHK dandach_h_Page_42.pro
defcd3a19293831ec0d0d993e0d7a653
102758b8428608b4d8a6b119898d612e0b31c79c
107823 F20101204_AAAXCN dandach_h_Page_21.jp2
5f7b27f05bc4907c2f83a7b0e41495ef
c77e7cdd51f2d920a8ba751e8045257ea76f98a0
21472 F20101204_AAAXMH dandach_h_Page_35.QC.jpg
a7fbb1839bf297d8bca8dc5f9abd5629
834d6d5c457fc52305a1d5e80be17c1686b36c2f
52416 F20101204_AAAXHL dandach_h_Page_43.pro
112fe04233a9a9b646db58895f0a2d38
1c5317cfb56dbc1ee5c77ce3db4edc9f27397873
79327 F20101204_AAAXCO dandach_h_Page_23.jp2
51e2a2bd718afc3598a4448697082ffa
de0aa01ed7ae447e2eab123de07c7e62c419ce90
24617 F20101204_AAAXMI dandach_h_Page_36.QC.jpg
f4e942724efba19c3c7f6c310a571a26
5d95536eb14536e8c4db19843694bac3d7081d95
41290 F20101204_AAAXHM dandach_h_Page_45.pro
e5530500bd056c0a370479e38536b7e7
23cf837d6c8c03ab3497fb8cc627a2f23fd870c3
502500 F20101204_AAAXCP dandach_h_Page_24.jp2
9ff81f88eb95cfa29221d6b9fc2c3234
75736d80f7d1874a439fcbbcc5c9721672f17dde
6633 F20101204_AAAXMJ dandach_h_Page_36thm.jpg
0e56d47910f886fb8221ea31b8750953
98343b1178289fdaee14bafddc34d80fcaa43699
50112 F20101204_AAAXHN dandach_h_Page_46.pro
ed0126193d60bc8c90bfb3b05dae46cc
708b6ffcefc29ce296d3972505f6fe22f48339f6
112185 F20101204_AAAXCQ dandach_h_Page_25.jp2
4fc680924ff0ec8d3c13673a038f2004
3895627fc693cbfd7d0cc4805d46be3ee59c8601
20170 F20101204_AAAXMK dandach_h_Page_37.QC.jpg
3909d4ce5d279fce9cecca11284b1a7a
85725a42447416c7a60439aa2b312fb6da0906e2
48926 F20101204_AAAXHO dandach_h_Page_47.pro
86635545e96c1923483ed30f64c9837b
8de3d11c8d2f08bc36feaa125161aac550fe2a4b
112254 F20101204_AAAXCR dandach_h_Page_26.jp2
e6f69250c9276bf28e50f621a8c7a658
4c793911c38656a046e2237eed32f681424b6af6
5732 F20101204_AAAXML dandach_h_Page_37thm.jpg
e433af904ff5c1bd8b18761e36d2e1cc
4d6bc7b763f7dcfce491e09c1f0c10371d9527fd
31450 F20101204_AAAXHP dandach_h_Page_48.pro
7c796012d5eef0ecd94657d5b2e65eed
5f0437f2ef5e4abea63f56aa04bb077571d10126
123502 F20101204_AAAXCS dandach_h_Page_27.jp2
19b8ae1f48330e709674e0f2896786e8
420c1ff4fa4169c56cd8244017966a4fa2eb387a
24703 F20101204_AAAXMM dandach_h_Page_38.QC.jpg
9bbe365308313cea9e829f224230598d
ff43e1ed6bcb4cad8fddd41f4020c02b9995e64d
24513 F20101204_AAAXHQ dandach_h_Page_49.pro
6ef260673e674dbcd35454b163af5247
59fa7f8a49354892171903379591b75b65e8fa52
45777 F20101204_AAAXCT dandach_h_Page_28.jp2
f7258d8e5e054a31d65d3c8bf4e203fa
f73a616eb20b146cde16bc6852a7c83444873ee3
6691 F20101204_AAAXMN dandach_h_Page_38thm.jpg
e91fdb924aa6a55d4cdf967f542466ad
7945346e2fe6b95e298b89243827f294de3ed73d
47791 F20101204_AAAXHR dandach_h_Page_50.pro
b54013b405568f77758b89becd01416d
b60169f29be39412b36c42626f2e8f5f7a7b2fae
23159 F20101204_AAAXMO dandach_h_Page_39.QC.jpg
f4d32e11e64d69a582d16ec5cb186857
a98c6ad5c5aa7003f0e30a0ac96d9e16c9012f57
51891 F20101204_AAAXHS dandach_h_Page_51.pro
559793cfc0cbf6a2148a07a7f6531172
b59a6ad701ebf7925e309186565a8be9f20cda06
101743 F20101204_AAAXCU dandach_h_Page_30.jp2
2514cdbfc98829e17469f9a0e682d321
1bb4df3547e3c6cff421f69036d7c8d662d9968e
6546 F20101204_AAAXMP dandach_h_Page_39thm.jpg
8ed636cc880e6f47c17802587ece1484
624dec3b8ca54bc99dec1ed0ab042c6d93189bb2
48663 F20101204_AAAXHT dandach_h_Page_52.pro
d3872bf81ea894a3606f6833eeb45719
378e539e697dcf28ae0fd58bf3dc83050cc99d75
1031552 F20101204_AAAXCV dandach_h_Page_31.jp2
2f26d57ad446ef6a0f9808f25b552d6c
c20016d3b7e8bce114ef26a5faef19a840f3c3c3
23789 F20101204_AAAXMQ dandach_h_Page_40.QC.jpg
9047eed2befa5258760f3db45cced69e
7e1eb6111b1489d3912a783151c68656b7f416e7
24814 F20101204_AAAXHU dandach_h_Page_53.pro
123e20ad211007eb1327bd2c731ab618
c0f3e8a846ba292f4b4c823cf3070e1ffc9a85d0
748990 F20101204_AAAXCW dandach_h_Page_32.jp2
11cd5366283a02a066678665f984f5a7
289a2d5a56f755df32947e84e5f527c64a290fca
6418 F20101204_AAAXMR dandach_h_Page_40thm.jpg
418af5145d1403276536298847e7e493
c3d4678844936ca351cb49f8ef4a68d66f4b78dc
22871 F20101204_AAAXHV dandach_h_Page_54.pro
6011e271ef465843e768d383709c4241
651f9b0c2fb240cf59ad957f209f8b7abb7c27b8
107234 F20101204_AAAXCX dandach_h_Page_33.jp2
c45cea5e2848d02c7de8df90689c7f2e
40c50be4c339d24afef9e23fd7d3d5f44e095882
13982 F20101204_AAAXMS dandach_h_Page_41.QC.jpg
9249c49bd039576ce74af8eeb6a5c536
18832c055518caa079d50113c4ea3549354fc8ac
57233 F20101204_AAAXHW dandach_h_Page_56.pro
012ded8d0cd3cf9f1d5ac00d39a1f738
1aee8bb2e40cef593f64c859e1dec15accaea8ff
13408 F20101204_AAAXAA dandach_h_Page_08.jpg
7e45da545635261232bea99465c707b3
bded9b43a5074c9349ece2e6ca092229a5794cea
94554 F20101204_AAAXCY dandach_h_Page_34.jp2
8bb0afde99daf5164a6a0fd4b18f8da8
d0b7d59eef6234623cd73a4ca3e48fe937a4e107
4054 F20101204_AAAXMT dandach_h_Page_41thm.jpg
543de23c678e98e174298b06c5076b13
71a9b86fbde410b1092f539ec15db8ea665ee080
45486 F20101204_AAAXHX dandach_h_Page_57.pro
177d3000fa83176f4f52c6a923c7dca8
18b2c615ac99022d3068bfc2bd9af99d215ab63d
70905 F20101204_AAAXAB dandach_h_Page_09.jpg
fbb8d9015be4b8a01248b074f2b3a714
4ac56a1df005128199156e8afc4d3eaa7b4a6634
95951 F20101204_AAAXCZ dandach_h_Page_35.jp2
5d7888f5b5714b2aef3078136114f5c4
c898553f028192fb31f19fb9f8e8ab364b212a12
24396 F20101204_AAAXMU dandach_h_Page_42.QC.jpg
4d51ba5bc862c0f0b047b9f5eacb3b85
a28a4bd7f38d83bcd33174176d4cbf18932a0dde
25817 F20101204_AAAXHY dandach_h_Page_58.pro
7bad03cbe52f7efd03d90a97ee4c20a2
27380234cf6fa90943466ccf9eef1316da63495e
69671 F20101204_AAAXAC dandach_h_Page_10.jpg
be85583a831f325ff54a999f1f0b5fba
c1a3fe85fcdc6a6e36c1e23cafc3fd6b73336517
6544 F20101204_AAAXMV dandach_h_Page_42thm.jpg
49d38e1723c5ea1a8bda6676cf9215e0
c6fde13b329180577f6e1a04fc193c89362fc01c
73117 F20101204_AAAXAD dandach_h_Page_11.jpg
8aa9d5f4989c95c5bda88ef2d5e036dd
b4b9ccd203a0172e322c835bbbe2943d33c61a21
23793 F20101204_AAAXMW dandach_h_Page_43.QC.jpg
3cf4b7f99a7c069ce989016d1666a973
2d2c24739e064ad3078bda8a860dd2ce4e44db62
F20101204_AAAXFA dandach_h_Page_32.tif
bcc120665d432f4cdc66079b9daadd14
c1eba4b8414ad410c03e54eaf70c421791f680bf
430 F20101204_AAAXHZ dandach_h_Page_01.txt
fd936136dcdcd0833452d3ae49e69fdb
0848c8c0e09738add958caf8de8eec8f76c70c97
35197 F20101204_AAAXAE dandach_h_Page_12.jpg
3a7173cb99d9ffb07a3953388bc049e0
0bb104d678d803ffee1b6a1cb49a861a7b494daa
6547 F20101204_AAAXMX dandach_h_Page_43thm.jpg
1132e93abece043a86dcad13cfb3241f
bfe536d134cbadd3a0fd90edee35401d22310252
F20101204_AAAXFB dandach_h_Page_33.tif
e81dc95e25270b818121366205a3823a
1c079b2dbe7ef41daf5a11c340a184f5cc460fe6
70596 F20101204_AAAXAF dandach_h_Page_13.jpg
c89b6abd689c6de7095c967985b048cf
132ed40fb8c882a4340c7019730c77642cb4fd6b
24331 F20101204_AAAXMY dandach_h_Page_44.QC.jpg
49a33d6ac558cec673d52b1dc5529b1e
f651340084cba5a6ad75e40dcacb59ba12390cc8
F20101204_AAAXFC dandach_h_Page_34.tif
9abecb84e60206ea3cad65df24f0c151
64abb780caf009b1f3a578b96e1815e817a55215
73007 F20101204_AAAXAG dandach_h_Page_14.jpg
fb36451aa34201a4e7d3aff2dd4b9f31
a4861a1174d8d3352aa7690fafa1d68e28189b4d
3134 F20101204_AAAXKA dandach_h_Page_02.QC.jpg
a05de5707c27e159fcd724f53558e889
e07a3a3ba6444eede850c08788e46000a5259a0c
6597 F20101204_AAAXMZ dandach_h_Page_44thm.jpg
ddbd9d8fa154c79a45176b6fc83cf0c3
7ff9174619e54b17f4a9c5ff93eedc82c5a4a2a4
F20101204_AAAXFD dandach_h_Page_35.tif
45665b57ad8de97236ccaf3a0f9198f0
345fe27ab984c35ed3bbe12e8a013623d5caaeb0
24042 F20101204_AAAWYH dandach_h_Page_29.QC.jpg
6bffb132e7059aa19eab1fadec3505e1
47e9b16110693b60ce8c91d5d22ac7b40e3af721
70795 F20101204_AAAXAH dandach_h_Page_15.jpg
671ed6ec78edec8b1ec3229a7799eec5
4dc80ff719638e9915a6a52fb7e2388da40c62ac
1337 F20101204_AAAXKB dandach_h_Page_02thm.jpg
18d93875ae8d9df38cb85ebed20b4630
b95979d850d0e71f3613846ee8169050a86e084c
F20101204_AAAXFE dandach_h_Page_36.tif
1a5ca4a755d58cd18ae8ba9b54711dcd
61665e33fce2d0fa4b6584b68a4f639b480827f8
51177 F20101204_AAAWYI dandach_h_Page_40.pro
96c297acf77972b9cc7fa958d1721083
e993df06f9c476ab44be2fff57c47496e1330fc5
71061 F20101204_AAAXAI dandach_h_Page_16.jpg
e29ad18bbb3824e7edfd9007e7126528
c57685879cc8295e2a0c7fd94e379179da751378
3726 F20101204_AAAXKC dandach_h_Page_03.QC.jpg
ddb370e38efe30335b036af5bf8e8e8d
18fc08220451ab533e40671f8671271cc6b3824c
F20101204_AAAXFF dandach_h_Page_37.tif
235a802c5a7c874ef6eb523e13267a13
5ac0f37f30add07b0a05d6243dd91ad3b859dc6f
6002 F20101204_AAAWYJ dandach_h_Page_35thm.jpg
f9c22652e2f7236004ae420021353d74
850bd3e4db4182d7dcbd8ce86fb99ce9da3b0343
70294 F20101204_AAAXAJ dandach_h_Page_17.jpg
ca87d02ef31f7f0ca3f91afcf0eabe59
402e966d288cd1bb4b1c5b98e1d7d1ebbfa75afb
1582 F20101204_AAAXKD dandach_h_Page_03thm.jpg
3cc67659bc5a943519099c1513b87aec
f6fb9168cd5800c9375615de7de94d93b8a3f396
F20101204_AAAXFG dandach_h_Page_38.tif
e45b10c1e1b5bcb8659543f36ef6dbc2
dceef99b3a041849db3a69e185fee722ccd430fb
66750 F20101204_AAAWYK dandach_h_Page_07.jpg
708652aca7b3a95d55c842ceab16130c
581daf23a3c495eb097973a835f02ada3e22b3bf
76041 F20101204_AAAXAK dandach_h_Page_18.jpg
c3aa74d542b51f246f58e7fd2973e331
d82a7d17b777356541c0b9b9468b4bc9f409c939
24575 F20101204_AAAXKE dandach_h_Page_04.QC.jpg
8554b7db715749ae79546a8dd058aaec
ad8f60d3c80045032ea62089cfccc623a0c71445
F20101204_AAAXFH dandach_h_Page_40.tif
9f3ad79f6b7760b556b3433dbe8e3328
5d1f4c0f8f7d19f23eeaffa1ad9d0c9bb5013a82
72850 F20101204_AAAXAL dandach_h_Page_19.jpg
d32fa0309b3854d4c0d2a6ca6231dedf
6c636bd8438dce7f769f22cc6407cbe925b2a35f
6256 F20101204_AAAXKF dandach_h_Page_04thm.jpg
89da746b3e8c309a483a8afda49f7408
078c5489f5fc46ca1e53813cd0ba3e39f139f9c6
F20101204_AAAXFI dandach_h_Page_41.tif
375b9e22b1841da2cfc6c0b34dbebdac
c92a1ad2910d4c8032a6b452c0a48c92cca184cd
8422 F20101204_AAAWYL dandach_h_Page_06.QC.jpg
1c2ae74ac568ac851a4e16104025d800
0763a9b02bcba2a82918c586ba1e81438b9a2c88
70022 F20101204_AAAXAM dandach_h_Page_20.jpg
6d94f40bba516d7197a63d3ccc17ff89
5660ef83261f9b8d27842e18341af836b23251b8
9042 F20101204_AAAXKG dandach_h_Page_05.QC.jpg
041074e295e3b549bd449b88ea03bd59
10f6b49b588d7a4b920231657b626d916ea496e1
F20101204_AAAXFJ dandach_h_Page_42.tif
0a2b373a86d4683ac9980694f5de5dee
b737d94a6ffa3b0e147d26113aab63fa865bbc6d
1791 F20101204_AAAWYM dandach_h_Page_34.txt
80ebf4f431d584bba949f6b2fd81d769
8dd34b568d4c0ca01a94c3ad9241e79709d6ecef
71701 F20101204_AAAXAN dandach_h_Page_21.jpg
a2b5c0b65348f3d0e9755853645fe45d
fe3f6a8b01c05f32abe196b76c372ec3101057c8
2705 F20101204_AAAXKH dandach_h_Page_05thm.jpg
8f4b52527fd4f4ad987d0e68e639d86f
97783370052ab27fa2631d8aacc0bd1bcdc993b9
F20101204_AAAXFK dandach_h_Page_43.tif
dff441d1a90f1262861231f6f5010a95
a4f292520d561ecb1ba5977bee48a96f17d4cc62
54664 F20101204_AAAXAO dandach_h_Page_23.jpg
504a063b85a76c49f3941ad00739205e
271cf1555b7d9d917358d5d72687190fca11160e
2679 F20101204_AAAXKI dandach_h_Page_06thm.jpg
981cae41f89eee0222a6cfbd66d338c1
361bf366db35b45715b207d6043d668733f6270b
F20101204_AAAXFL dandach_h_Page_44.tif
be9d659db184338fac1f5c009e6af122
18c3274d03ffe7aa23540bc869f8711421c8c9a3
2115 F20101204_AAAWYN dandach_h_Page_18.txt
ecc9d6c4349769367507abb31b120bbd
84c132436a94f5b4870ee5998d7721bde863262f
37963 F20101204_AAAXAP dandach_h_Page_24.jpg
077ca654c2651d047456915be88d764f
3256b33f434e79a2ace68d0254ff440310d8f479
5949 F20101204_AAAXKJ dandach_h_Page_07thm.jpg
53f979aea6a4deca0a100c905f226f50
4ef941e1cdf165fb98bda8062cffcbcde6b084c3
F20101204_AAAXFM dandach_h_Page_45.tif
53f0c9aac2a01b002265135b20e2ddac
579b4085c7cc2963ebe909e8d84ceeb290b047d7
2184 F20101204_AAAWYO dandach_h_Page_44.txt
3ada18e6bd6b6604a444f83a353263c5
4ca123dadbcbb700e23fa281df46c5559412006a
75661 F20101204_AAAXAQ dandach_h_Page_25.jpg
bd3e6cad1b8e549a33e8001853c9e936
3860ba22f55d26714d0c4895a2e8bbb5d1eb51e2
4445 F20101204_AAAXKK dandach_h_Page_08.QC.jpg
4f0560aeeaf7ea255e579f3853f2e119
324658e5fd683b6ce0ce676ab7aacab81c71e615
F20101204_AAAXFN dandach_h_Page_46.tif
ea6a79b9c5371d3e01322af1d85eb5a3
f5c1a13b3fb5925dbbeeabef7498798fad2906b4
F20101204_AAAWYP dandach_h_Page_48.tif
b9d40d0ceeee0f82f06d4d80562548aa
be1c48a1a021687af80df4235eb2644d30ad37d3
74805 F20101204_AAAXAR dandach_h_Page_26.jpg
dc46946c051c6eefe5563fa6bc89400e
e94be0c796cdb63c7ff1e3d8d330885225f09e76
1638 F20101204_AAAXKL dandach_h_Page_08thm.jpg
2b1df9c9c3fe30769b9321b1c0e423ff
e435772d9aed28091e31fe4f4eb0ffdc34c61aed
F20101204_AAAXFO dandach_h_Page_47.tif
098609979c9ce5f64d20deac4eeda7b3
3049c53f60d04ed09a5e76ee0e3c537b1e9ccbe3
1401 F20101204_AAAWYQ dandach_h_Page_48.txt
6c08fe18138db0097e278051e7579ac2
5ab1d146c23ae223a63b94439b8e2a347e1ee824
22889 F20101204_AAAXKM dandach_h_Page_09.QC.jpg
a2705632ab9113b4275ed637921e0ca1
59b0d2755c8ba03e3469d014f0445311dff83dbc
F20101204_AAAXFP dandach_h_Page_50.tif
0cc3f5a45d6e9693789c49b5051e6ec3
2bdff2d987a13a83ff19513257086df1d2c008e6
75268 F20101204_AAAWYR dandach_h_Page_36.jpg
d78b8d875b6c2206fff39102445cf608
fff11661068385a800e3e24fd17b91cf98a88a3e
82737 F20101204_AAAXAS dandach_h_Page_27.jpg
79d2930eb642bf3b09462c2fb4958b17
8acd1ef7e68f6b52d0338626f00ec8d633b67e2c
6401 F20101204_AAAXKN dandach_h_Page_09thm.jpg
71b1067a1e96161961041da9e084f3f0
f81afcb7ee928e64e7c1f193c72b1ccece0ac081
F20101204_AAAXFQ dandach_h_Page_51.tif
701afdc0ee60a22d48633d608e5e1116
2a80d54315308285660e217c3908477c93aa0064
56517 F20101204_AAAWYS dandach_h_Page_55.pro
dd3dcfe06730f995615053495c8674bb
3362fefc08bfcd20ae0fbe02b7d34777f7192a3b
33689 F20101204_AAAXAT dandach_h_Page_28.jpg
a3527c69fd53fac3dda06ec15cd6cce2
e9243d26c4b2431e2e7326ce019cb5317c69660b
22815 F20101204_AAAXKO dandach_h_Page_10.QC.jpg
de28662376457247f91aa7f7ab1cd938
22a32ef8c15c6a93996cae23752abaff3d1500dd
F20101204_AAAXFR dandach_h_Page_52.tif
500dfa46a8b774490018aacc4c4bd71c
a9b95d00fec6d773ba15d44954f45636f74af264
2305 F20101204_AAAWYT dandach_h_Page_56.txt
4aef190961bb501a8081b67b8d25632d
5d79e9aa5fa202fdd2eaf7c491692da6af3359bb
6255 F20101204_AAAXKP dandach_h_Page_10thm.jpg
85802947bfb20a1c57560e56fea8e82a
7205a1144a3cf70767293e3bfbca856a9f1e2b06
F20101204_AAAXFS dandach_h_Page_54.tif
29cab440362797223b0208ed5b63320a
6a9d7e75491e748b63c6c4a785375bc12954cc60
6631 F20101204_AAAWYU dandach_h_Page_26thm.jpg
ddcd5c6005e8732f0123e5402b899d57
7744beeb3ff7edde11bd11ccf4955b6b0e440d9d
73930 F20101204_AAAXAU dandach_h_Page_29.jpg
51583feba1ecc1191a4587cba3b2eb1d
dbce752554ef99cc91fb8ee67d1650cd64fbbdd3
22551 F20101204_AAAXKQ dandach_h_Page_11.QC.jpg
3db2abea9f9206858944e99209a9c83e
7653b5874d34e9fd5ef95d88a49ee0be29776f60
F20101204_AAAXFT dandach_h_Page_55.tif
c4ac1bd647dfd20ee1df93e2286a7cc4
ff113d191f6d1440a6b929a79f991ec183b02c5f
F20101204_AAAWYV dandach_h_Page_12.tif
42ccc3e0ef87cd5d6eb29e6eb7cf8544
dc2a4e66686f464d28d78ef9bbc699a51a05bd1c
67517 F20101204_AAAXAV dandach_h_Page_30.jpg
a5ba2463490f15d61f6f49af3c8ed826
f0aba09ba96869324697a593c685cad115fa1894
6259 F20101204_AAAXKR dandach_h_Page_11thm.jpg
899e4739a9cc05c054852e264110bd2b
a9eae7544dd5fe7d67b3e1822a36b27437641edc
F20101204_AAAXFU dandach_h_Page_56.tif
98a59340a8e15ac654be65622a641f8e
899dab446dc065430b97657d42651f817f3d1214
3962 F20101204_AAAWYW dandach_h_Page_54thm.jpg
c4c9cbb98ccb511ba33946d467c543b6
bc71ed70944ba0064bc22897632d43035a80f409
73456 F20101204_AAAXAW dandach_h_Page_31.jpg
52db812d50ae3f623863dafc19841f45
3124cdbfee548a4ebb5173c07fa5ba94135d47bf
11141 F20101204_AAAXKS dandach_h_Page_12.QC.jpg
a430c026502892a8362815988a5b080e
393cf2161c2a4eb3c1df362453e230a0e3b2b90b
F20101204_AAAXFV dandach_h_Page_57.tif
de8432cdcc486fed3e40365e7fc98c7f
2cb74a062b1fb582b110ff965730b6e165ac5e02
39047 F20101204_AAAWYX dandach_h_Page_22.jpg
15f0013cd3b258c232d99b7192e37c5c
73d588e9274c9d9636e56b6cafe0250b2f670d32
71349 F20101204_AAAXAX dandach_h_Page_33.jpg
e74f3cd2ea929834a693cc3b13ba0877
566ad694f37bb27c8a3a2eccbe09bdfc80f50692
23470 F20101204_AAAXKT dandach_h_Page_13.QC.jpg
0f6940af0a9951b64fa9ffc5833538b2
e982e53fac258b756c77daa8604f4fbc9e5495a4
F20101204_AAAXFW dandach_h_Page_58.tif
c6b2ff6352b816c464bc36b1d6a8f30b
623f647c964667c0e80fddc3ad860d65f0f340e1
20757 F20101204_AAAWYY dandach_h_Page_07.QC.jpg
58d25b132dcc9a728acb9ec3abe8ac1f
cc64a4406f37117e85b5e563881f866b707a6510
64450 F20101204_AAAXAY dandach_h_Page_34.jpg
a3c80e0275303d1e40c683150bc6af09
73bb75fa7ba3b54f8a1b9850d770b3001afbcf02
23298 F20101204_AAAXKU dandach_h_Page_14.QC.jpg
d1c2655d9d1a35522e5d61f923c8b35a
212f0f828a3fca6d4192d42134f3b0bb7e1c0584
68244 F20101204_AAAWYZ dandach_h_Page_57.jpg
13d1243a7705671ddc1898cd1f40e10d
6cdb9d2137f625a7c835f3a9cf67eb2e11f13c75
65595 F20101204_AAAXAZ dandach_h_Page_35.jpg
2fff8eb2b8450b4282937e95abf2a6a1
314dbae76618bdec52d51eb86041ca0cbdb410d9
6432 F20101204_AAAXKV dandach_h_Page_14thm.jpg
dce8d24aaf088ede2059a81e17dbe97b
0013e5ea69047edfebe292c995a65fd33beb7dc2
8123 F20101204_AAAXFX dandach_h_Page_01.pro
5afe40916b12561142813e078a87cd83
ab30239fee309527eeb3dd7bf25ceae69dcc438b
22957 F20101204_AAAXKW dandach_h_Page_15.QC.jpg
a6fe264e44ad3f46c724d2ae7b0d7f18
6e08f626f209eb67c33a7506293b60a58b683056
112731 F20101204_AAAXDA dandach_h_Page_36.jp2
b234d3e57cac0b92bafae1de4ecdaa20
c525636cff3a1f88cf043b35eaa40e65ba5a33a8
1016 F20101204_AAAXFY dandach_h_Page_02.pro
7ce01d430ab769813e49b77d6d9578db
37d6ec1192fb9446acee0e74075158600d1c72f2
6385 F20101204_AAAXKX dandach_h_Page_15thm.jpg
a822bdb2514a5c45cc7f2fcaa47fc492
f26a812fab6eb11a48474b14352ebe818e8bdd96
95924 F20101204_AAAXDB dandach_h_Page_37.jp2
427c7ca3385c46f841b5dc7a0d2145e5
61fd2fe79bc6a139a3f5ae49a56531e0516580c4
2652 F20101204_AAAXFZ dandach_h_Page_03.pro
2c4acd23cce360110fff89a693de5aa4
557925d6776923237dd99f015cb2fa958e46964b
22959 F20101204_AAAXKY dandach_h_Page_16.QC.jpg
fa6c4267fec6be11f3eb038962a70c64
f46f76dfa9244b014e4296150093f0fd2a105447
112860 F20101204_AAAXDC dandach_h_Page_38.jp2
f68a0b809b20debba48780ffd77a26e7
2fea258411dd92aa3426ba361343b3986b83aff6
6455 F20101204_AAAXKZ dandach_h_Page_16thm.jpg
944ed69b87cb9a7531edf4b31af0301c
990d8c655db20c37237f6cdac492c0468a713389
105332 F20101204_AAAXDD dandach_h_Page_39.jp2
86d4e72c5ac76bfbb352c98cb52e915c
5f3616626bb1b1138656a3e67ac1d6bcf52aedcf
95 F20101204_AAAXIA dandach_h_Page_02.txt
b79dcf67dfd2ff6d8070143f928c3206
dcd415bb25ade48fdd763e7aa9f2c817b7d6ace7
106547 F20101204_AAAXDE dandach_h_Page_40.jp2
3dc1ab77c33ad6d5da54e6140b9c8570
784c53300ac3ebc103a6fe6fa3a659c0120dbff0
3584 F20101204_AAAXIB dandach_h_Page_04.txt
9d4effc15bb9c3fb5eeb04ef97062db8
24e85aaf9c275e31714a1f8f78869299fcb8b86e
111215 F20101204_AAAXDF dandach_h_Page_42.jp2
e81efb47416c63d686c099200f1c3a23
62a70c27a9126e541e1d882f8a4ac6ef0971e4e4
1129 F20101204_AAAXIC dandach_h_Page_05.txt
5462e256e4c8d635827eec7e7f6fd15c
00d693d5269327abcb89087a6399e4c145948ba5
109041 F20101204_AAAXDG dandach_h_Page_43.jp2
adfc497e8ca85047b9d52c286e555a58
16769f663e9b690d3e165bc9a8ee97504e072e66
21694 F20101204_AAAXNA dandach_h_Page_45.QC.jpg
6332e93a5be3a17eb091247d8a6ef18f
2624fcd1a04fb540e1f7acee733a42f37cb0eecd
809 F20101204_AAAXID dandach_h_Page_06.txt
1253fd4d68e31771b4e75469bcf32e24
8ee98971cc97b9e20d9cf6a0b920cea9d8fdd1f9
112489 F20101204_AAAXDH dandach_h_Page_44.jp2
f2fb072fd4f0bb1bdefa62d518f46d1b
e05d3061c946af03db3f98c0b707e2d236eb531e
6269 F20101204_AAAXNB dandach_h_Page_45thm.jpg
693be6339fe04ec11fea75fd77199bf3
93695b49f213f4f7ebf1e6c06b3e56edd10400ed
1926 F20101204_AAAXIE dandach_h_Page_07.txt
b74d8cf3d4ad978b434a1bae3c086738
71caa9d713fa84257a0c4d1c40b07754aee84335
946574 F20101204_AAAXDI dandach_h_Page_45.jp2
e9abd32ce6995b53e697ec00b3687459
1e45074a8ad0c511e0d91a003e1cad2f682aa993
23473 F20101204_AAAXNC dandach_h_Page_46.QC.jpg
6c1c5197b4e36895948c805e5484495e
a81290e9607a03ced7ee27d349eb8147c881a554
151 F20101204_AAAXIF dandach_h_Page_08.txt
6ac8b4691c8e8659aa60695c10a91803
e42a501584903f2eca1817bc6e78294ba678811a
105034 F20101204_AAAXDJ dandach_h_Page_46.jp2
02a3c5faa647c498016734a381579c8d
e0294070ad8d115e3d29f0945a511171a7f75d05
F20101204_AAAXND dandach_h_Page_46thm.jpg
fca9cd9fd86a5eb7b4da449de104e7e7
df352dbd1c731ff40d6bee80fc119b8f6e43fa29
2039 F20101204_AAAXIG dandach_h_Page_09.txt
203d5d6571eb0737d643fdfd7faa04c8
2fadfe276f195bff7f9a8a622ed9c85b46e98018
103943 F20101204_AAAXDK dandach_h_Page_47.jp2
83c3c37ff0717b8f410c2f0d2e90248f
dfd2c87edd034cc181f914eb4582902ef7e5247f
23024 F20101204_AAAXNE dandach_h_Page_47.QC.jpg
8b602dbcaa343874badbb83af0e5c6b8
de004a689bd5f263a0d999bb0466bca9e155a738
1884 F20101204_AAAXIH dandach_h_Page_10.txt
672093a0b6670ba543472a1925e21d0d
daac6f009863997db4503529e84ba6ee1289d52d
843405 F20101204_AAAXDL dandach_h_Page_48.jp2
43496b6d3a3965242fddc138aac837b7
0338740a3e6f53cd9f081a670468da640f5d1a43
6139 F20101204_AAAXNF dandach_h_Page_47thm.jpg
49cfc3c9b5bb43dfd6671646d9060cef
2df42f4ddb372d346651ee3e17d478f94d4fbd1a
2121 F20101204_AAAXII dandach_h_Page_11.txt
f8fcd464b5e8081750e82ec96134b76a
70832bbc5f9f820f93280156fdb4fa4279ee4bd4
54329 F20101204_AAAXDM dandach_h_Page_49.jp2
e2501c64104ced9a8316e4f6a673316e
4610539c5912f8106e253a7852191eb30f2cb760
17672 F20101204_AAAXNG dandach_h_Page_48.QC.jpg
851037ed77b178d578e03f1a7c72176d
ce75de25c36ad5088912ea4ab63d37b0dc17633a
1990 F20101204_AAAXIJ dandach_h_Page_13.txt
1eb23a1ccbf378d58f779e644f9bb900
c312dff5c277f83ec47fe80e59eea90ecdc92940
100566 F20101204_AAAXDN dandach_h_Page_50.jp2
3888202058a5432808bf9afa7797d2a2
28cbe0f56c1d81dd6e6eba9a1d350e5688ae176a
4944 F20101204_AAAXNH dandach_h_Page_48thm.jpg
4e18769240aa6ba65889121b2633cb48
cc388bf7127f5b19ab3d020f79d38af56d70905a
2038 F20101204_AAAXIK dandach_h_Page_14.txt
f909ea0f96ee96208040ce95cb56b03f
062092c3563c2dd4d7c48891d208b6f780dc6058
110418 F20101204_AAAXDO dandach_h_Page_51.jp2
6cc133b3a24d14ec5c9f012a115ea952
82d61d1667d6663bedf3a926427b2ad16c04888a
13238 F20101204_AAAXNI dandach_h_Page_49.QC.jpg
9a3db3a33f178ec2b143fc0a0114c9a0
df8f777ed20d55cb1d98ad023e95ab4e57eaef7d
1930 F20101204_AAAXIL dandach_h_Page_15.txt
77cccd6724c06b38b8b6217b73ed51b2
4909cf1df16ad74e85274115488e3d2c3618914c
104232 F20101204_AAAXDP dandach_h_Page_52.jp2
a2f34dd895dcadc68e5cd17bbfe0848c
be4146827e74d00398474613b0231ea1582b4637
3739 F20101204_AAAXNJ dandach_h_Page_49thm.jpg
7ae7f0ed6a6df72605e8a4ff4cc27a92
3c2d2cd8e3d1478175aa292d7717b8f2f5be4d5d
F20101204_AAAXIM dandach_h_Page_16.txt
e0e461e02f965f4cd3bf36065bcdecc2
b2245086ac211b82161a451e63a64b260e97f2a2
55114 F20101204_AAAXDQ dandach_h_Page_53.jp2
11d8c83e7b75ab713e49bfb47ec128db
49437b70adce77550a272b3910565201c6e9e9c7
21847 F20101204_AAAXNK dandach_h_Page_50.QC.jpg
2db6006f88378e1e22dc56ff67be2448
54be4755ac9f55fa0d4898bb56019a4b8fdea374
1951 F20101204_AAAXIN dandach_h_Page_17.txt
4790f314bfe638a374fa899e50f8fd26
ff0d1fc0611c0e552081f1042385c9deb269166c
F20101204_AAAXNL dandach_h_Page_50thm.jpg
3ca8508180526cdeab321674227642e3
6f74559532b9401c31c61caddc2fda031d9dd8cf
2042 F20101204_AAAXIO dandach_h_Page_19.txt
76e7cb5bbf46f19ad2cb78d071f16caf
7767d88eedd4e55f844480a2dd131555024f9229
51821 F20101204_AAAXDR dandach_h_Page_54.jp2
a39480c95c121d34148b63ab97ede6d4
d800064f5aab68290458d96e202dff28677ce702
24045 F20101204_AAAXNM dandach_h_Page_51.QC.jpg
ee4ba55251c3a37906157e4b4420471d
e00c628dcc82856b04e19e2e4d19fc852c65b777
1948 F20101204_AAAXIP dandach_h_Page_20.txt
cb5f520661be7281296b69a1e8270ada
f8a59eaec537d1fc15676d7287eea14653e8eeb5
122016 F20101204_AAAXDS dandach_h_Page_55.jp2
50d2af488e8cc1fd4d5d43a5c15ba698
07e492529c2a76f65d98b1d7488ab9ddbd869ac4
6472 F20101204_AAAXNN dandach_h_Page_51thm.jpg
9114170d65bd90ee60d6198d94d219c8
ae0031fa45390a5532d987dd752f788a72cf35fa
1970 F20101204_AAAXIQ dandach_h_Page_21.txt
cc864e258b35575b4e9e77c36c4be115
0fce236913d2a4d82ba0f0098809b2d0fe6991b5
126127 F20101204_AAAXDT dandach_h_Page_56.jp2
49e1e6e15cc641701f795228c870767d
32d8371c60303736bd1ad4865ceaa7269bbed588