MATHEMATICAL PROGRAMMING SOCIETY NEWSLETTER FEBRUARY 1988
Chairman's Report to
1987 has been a busy year in the affairs of the Mathe-
matical Programming Society. Fueling these affairs has been
a continuing discussion on how to extend the interests and
benefits of membership in the Society to embrace the
developers and users of the models and methods of mathe-
matical programming as well as the theoreticians, and to
englobe a more thorough mix of nationalities. While this
remains the goal, several steps have already been taken, most
notably the reorganization of the journal and the fact that our
next Symposium will be held in Tokyo.
Beginning in 1988 the journal Mathematical Program-
ming will be expanded and reorganized in two parts, Series A
and Series B, and the Studies will be discontinued. Series A
will carry on the function of the journal. It will be published
in two volumes per year (vols. 40 and 41 in 1988) of three
issues each, with each issue comprising about 120 pages and
appearing every two months. Michael J. Todd will continue
to be the Editor-in-Chief. Reducing the traditional journal
role from three volumes back to two has seemed for some
time to be a necessary step.
Series B, subsuming the function of the Studies, will
permit the timely publication of issues focused on particular
subjects and responding to the interests of the mathematical
programming community broadly conceived. It will provide
a means for exploring new directions in applications and
computational practice as well as theory. It will appear in
one volume per year of some 520 pages (vol. 42 in 1988),
printed in three issues, appearing separately throughout the
year. William R. Pulleyblank will be the Editor-in-Chief. It is
expected that research symposia will often serve as the foci of
issues of Series B so special provisions have been made with
the publisher that enable individual issues to be published in
book form for distribution as a "proceedings" volume or in
other guises for other purposes.
Every member of the Society will receive both Series A
and Series B. This represents a net increase in number of
pages of some 15% and is made possible by an agreement
concluded with North-Holland this spring after several
months of intense negotiation. The officers of the Society are
continues, page two
'88 Symposium Report:
Programming in Japan
In this short report, I would like to
introduce some features of the Japanese
Mathematical Programming community
and hope it will help the '88 MPS visitors
understand our state of the art.
Like other countries, the MP activities
in Japan began as one branch of the
Operations Research Society (ORSJ) which
celebrated its thirtieth anniversary last year.
In 1956, S. Moriguti introduced Dantzig's
simplex method to Japan. Since then LP,
network LP, NLP, combinatorial program-
ming and so forth have been introduced,
studied and applied extensively in our
In 1960, M. Iri published his original
work on network flow theory. If we call Iri
et al. the first generation, we have the
continues, page three ..
REPORT: MPS'88 & JAPAN 3-4
CONFERENCE NOTES 5
TECHNICAL REPORTS &
WORKING PAPERS 6-7
JOURNALS & STUDIES 7
BOOK REVIEWS 8-11
----. --~---- ----- ----
OP T I M A number twenty-three
satisfied that this is an excellent contract.
The membership should know that an
ongoing clause stipulates that if an institu-
tional subscription is due to the promo-
tional efforts of the Society or one of its
members then the Society enjoys a continu-
ing bonus of 10% of North-Holland's net
receipts from this subscription: so if anyone
has reason to believe he or she persuaded
an institution to subscribe please inform the
13th International Symposium-
It was my privilege to visit Tokyo in
November 1987 as Chairman of the Society,
to meet there with Masao Iri, Chairman,
Kaoru Tone and Hiroshi Konno, Vice-
Chairmen, and the members of their
Organizing Committee, and to visit the site
of the Symposium, Chuo University. I was
very impressed by the enthusiasm and
efficient organization of the Committee,
with each of the many participants assum-
ing one or another specific responsibility.
The facilities at Chuo University are
excellent. All lecture halls are concentrated
in one building and each is amply equipped
with projectors and blackboards. If good
organization, good taste and effort are any
measure this Symposium will be a huge
The chief problem is cost due to the
level of the yen. The Organizing Commit-
tee is working hard to collect extra funds
but in an increasingly difficult competitive
environment. They are anxious that it be
remembered by all that the registration fee
of Y 26 000 was worth $100 in August 1985
when it was formally decided to hold the
1988 Symposium in Tokyo, and the fee
includes a book containing tutorial and
survey lectures. My observation is that
Tokyo hotels (with $1 worth 137 yen when I
was there) are reasonable compared with
hotels in the capital cities of Europe and
The increases in dues for 1988 over
1987 are some 15% or less in all currencies
except the U.S. dollar. The increase from
$45 to $55 determined several months ago
is an increase of 22%, this exception
obviously a result of the fall in the value of
the dollar. The increases are otherwise
explained by (i) an increase of 15% in the
number of published pages received by
members and (ii) an increase of over 25% in
the fees paid to the International Statistical
Institute for secretariat services. Located at
the Hague, ISI's cost are in Dutch guilders
and they understandably insisted on
changing our contract from one defined in
U.S. dollars in 1983 to the local currency.
Executive Committee. Energetically
chaired by Jan Karel Lenstra since Septem-
ber 1986, this committee is the body that
manages Society affairs according to
policies set by the Council.
COAL-Committee on Algorithms. Jan
Telgen, Chairman. Members: P. T. Boggs,
D. M. Gay, J. K. Ho, K. L. Hoffman, R. H. F.
Jackson, G. Mitra, J. B. Orlin, D. R. Rardin,
K. Schittkowski, R. B. Schnabel, W. R.
Stewart, P. L. Toint, S. W. Wallace.
COAL-Newsletter. Robert R. Meyer,
Editor; Jens Clausen, Co-Editor.
Publications Committee. Laurence A.
Symposium Advisory Committee. Martin
Fulkerson Prize Committee. Manfred
Padberg, Chairman, Martin Gr6tschel and
Dantzig Prize Committee. Olvi L. Man-
gasarian, Chairman, Katta Murty, George
Nemhauser and Margaret Wright.
Membership Committee. Karla L.
COSP-Committee on Stochastic Program-
ming. Andras Prekopa, Chairman, John
Birge, Jitka Dupacova, Juri Ermoliev, Peter
Kall, Yves Smeers, Roger Wets and William
Orchard-Hays Prize Committee. John A.
Tomlin, Chairman, Richard H. F. Jackson,
Michael J. D. Powell and Michael A.
A. W. Tucker Prize Committee. Robert G.
Bland, Chairman, Hrold W. Kuhn, Alan C.
Tucker and Lawrence A. Wolsey.
OPTIMA-The Mathematical Programming
Society Newsletter. Donald W. Hearn,
Editor, and Achim Bachem, Associate
To date four groups have proposed
hosting and organizing the 14th Interna-
tional Symposium on Mathematical
Programming to be held in 1991. The sites
and group leaders are: (i) Amsterdam, The
Netherlands, Jan Karel Lenstra, Alexander
Rinnooy Kan and Lex Schrijver; (ii) Lau-
sanne, Switzerland, Thomas M. Liebling
and Dominique de Werra; (iii) Rio de
Janeiro, Brazil, Celso Carneiro Ribeiro; and
(iv) Stresa (close to Milano), Italy, R.
Maffioli. In view of the sparse membership
in South America the possibility of Rio is
being postponed to some future date. A
firm decision will be made in August, 1988
The Brazilians are not, however, losing
any time. Ribeiro is organizing an interna-
tional workshop on Mathematical Program-
ming to be held at the Catholic University
of Rio de Janeiro, October 10 to 15, 1988.
This conference will undoubtedly be
sponsored by the Society.
A major membership drive is about t
be launched under the direction of Karla
Hoffman together with a broad based
The COSP Committee is planning to
organize the 5th International Conference
on Stochastic Programming in Ann Arbor,
Michigan in 1989.
The COAL Committee, through its
Chairman, Jan Telgen, has voiced to the
Executive Committee its concern about the
direction in which the Society is evolving as
reflected in the papers published by the
journal which they view as increasingly
theoretical in quality and quantity. They
have in writing formally encouraged the
Council to consider this view and to decide
whether it wishes to do something about it.
In concluding this report I would like
to encourage all members to seek to use the
Society-their Society-to further their
professional goals in ways that construc-
tively promote the area of mathematical
programming. After all, the Society exists
only to serve its members, so members
should be inventive in discovering needs
that can be served.
Michel L. Balinski
~ ~~ ~ - ~ _~ ~--~I-_ ~I
second generation represented by the
studies in the complementarity and fixed
point theory by M. Kojima et al. And now
we have the third generation which has
entered the community with the interest in
Karmarkar's new LP algorithm. (It should
be noted that the MP community needs
young and brilliant talents to develop
theory, algorithms and applications but the
number of talented people are very limited
and many promising fields such as com-
puter sciences, superconductivity and bio-
sciences are sharing them. We need new
fertile subjects of interest to attract such
young people. Karmarkar's method is such
a rare and big chance.)
;apan Mathematical Program-
Regarding the forthcoming interna-
tional Symposium, we must note the
Japanese symposia held since 1980. At that
time, two MP research groups one in the
Kyoto area chaired by T. Ibaraki and the
'er in Tokyo chaired by myself, had been
/anized under ORSJ. The two groups
joined to hold the annual symposia of
which Iri has been acting as the general
chairman. The symposia usually consists of
four sessions organized by invited speakers:
sessions by the guest speakers, recent
advances in MP, MP applications and
special topics of the year.
I will list some of the titles from the
programs of the last three years, not to be
exhaustive, but just to show which subjects
are being studied in our country:
1985 Guests: Preprocessing of mixed
integer programming problems (E. L.
Johnson), On parallel processors design for
solving stochastic programs (R. Wets)
Advances in MP: Trust region in uncon-
strained optimization (Yamashita), An
algorithm to compute a sparse basis of the
null space (Kaneko), Solution methods for
variational inequalities (Fukushima), A
Multiplicative penalty function method for
LP (Iri and Imai)
Applications: LP applied to the planning
raw material purchasing (Hanai),
plications of optimization methods to
satellite communication systems for
effective use of orbit and frequency
spectrum (Mizuike and Ito), Production
planning software for wide variety of
products in polymer processes (Shikakura)
Polynomial Order Algorithms: Recent
topics on minimum cost flow algorithms
(Fujishige), Geometric search algorithms
and their applications (Asano), A hybrid
method for LP (Tone), Notes on improve-
ments of Karmarkar's algorithm (Kojima)
1986-Guests: Pathways to the optimal set in
LP (N. Megiddo), A class of algorithms for
sequential and parallel solution of algebraic
linear and nonlinear systems (E. Spedicato),
Applications: House reapportionment
(Kosiyama and Ichimori), Monthly produc-
tion scheduling of passenger cars (Ueda),
Analysis and evaluation of human knowl-
edge of power system operation (Komai
and Sakaguchi), Today and tomorrow of
Toyota production system (Kimura)
Production Control and Scheduling:
Scheduling in Toyota production system
(Kotani), Recent topics on scheduling
algorithms (Kawaguchi and Kyan), A
shortest path problem with visiting order
constraints (Suzuki and Nomura)
Advances in MP: Methods of fast auto-
matic differentiation and applications (Iri
and Kubota), The revised Karmarkar
algorithm and its computational experi-
ments (Kojima and Tone), Strong unimodu-
larity for matrices and hypergraphs (Ibaraki
1987-Guests: Notes on defining propor-
tionality between matrices (M. L. Balinski
and D. Demange), Optimization problem in
electric power system (Aoki)
Computers and MP: ID based cryptsystem
(Tsujii et al.), On message complexities of
distributed algorithms (Hagihara), Area-
efficient drawings of rectangular dual
graphs for VLSI floor-plan (Tani et al.)
Applications: New production & material
flow control system in steel work (Ueno et
al.), The production system of MAZDA
(Sasaki), Practical reactive power alloca-
tion/operation planning using successive
LP (Suzuki et al.)
Advances in MP: Centered Newton
method for LP and QP (Tanabe), Stationary
point problems and a path-following
algorithm (Yamamoto), Sparsity and block-
triangulation (Murota), Nondifferentiable
programming and semi-infinite program-
tion, networks, matroids
and submodular functions.
Fixed point algorithms,
stationary point problems.
T. Kobayashi Network programming.
T. Oyama Apportionment problem,
tion, network optimiza-
LP, NLP, semi-infinite pro-
gramming, application of
I ) AGE 30 P TIMAnmetenytreERUY19
~1. ~ II ~I~ s~- Ps. -~----~-~ ----~
0 P TI M A number twenty-three
Who's Who Around Tokyo
The 13th International Symposium on
MP will be held at Chuo University in the
center of Tokyo. The organizing committee
consists of faculties belonging to several
universities and institutes around Tokyo. I
will list their fields of interest so that
visitors can identify those sharing the same
University of Tokyo
M. Iri Computational aspects of
MP, applications of
optimization and algo-
K. Murota Combinatorial optimiza-
tion, numerical analysis.
K. Sugihara Computational geometry,
tion, solid modelling,
Tokyo Institute of Technology
K. Fukuda Matroid theory, combina-
M. Kojima LP, NLP, fixed point
theory, stability and
H. Konno IP, nonconvex optimiza-
tion, optimization in
finance & investment.
S. Mizuno New methods of LP,
U. Suzuki Applications of combina-
PAGE 4 0 P ubrtet-he ERAY18
'88 Symposium Report
I. Kaneko Decision making models,
networks and networking.
T. Asano Computational geometry,
MP and its applications,
queue network theory,
M. Ozawa numerical analysis,
S. Morito Interface between mathe-
matical scheduling theory
and practical shop
Science University of Tokyo
R. Hirabayashi NLP, IP.
H. Yabe NLP, nonlinear least
S. Shinoda Network optimization,
tion, applications to
R. Manabe Network programming,
applications in industry.
The Institute of Statistical
K. Tanabe NLP, interior point
method for LP from the
point of view of the
T. Tsuchiya NLP, LP, automatic differ-
Central Research Institute of Electric
T. Ohya Computational geometry,
MP in electric power
Will you find interest in MP activities in Japan? From my very personal point of view,
there is definite interest because MP is gradually becoming the method of choice for
promoting the efficiency of our industries even though they are not completely aware of
the fact. With the limited growth rate in recent days, what else can better plan their
efficiency? The industries include not only manufacturing but also services, marketing and
finance, reflecting the shift to so-called soft engineering in the Japanese economy.
Although I introduced only a part of MP activities in our country, you are sure to see
many interesting people from the four islands of Japan, all of whom welcome you enthusi-
Looking forward to seeing you soon in Tokyo! -K. Tone
Daily Expenses for Meals and Local Transportation in Tokyo
The 13th International Symposium on
Mathematical Programming will be held
August 29-September 2, 1988, in Tokyo.
However, it seems that the recent high rate
of exchange of Japanese Yen may cause
foreigners to hesitate to visit Japan. In
response to the Editor's request, I will try to
describe how much our daily life costs in
Costs of meals span a wide range, from
inexpensive to very very expensive. In
general, Japanese restaurants present in
their show windows samples of dishes
together with their prices. Restaurants
without such samples are usually very
expensive. Restaurants in hotels are also
relatively expensive. We can find less
expensive restaurants in almost every part
Many coffee shops set "morning-serve"
time typically from 8:00 to 10:00, during
which breakfast including bread, eggs,
vegetables and coffee, for example, is
offered for 400 500 Yen. Many restaurants
set "lunch-service" time typically from
11:30 to 13:30 or 14:00, during which lunch
is offered for 500 1,200 Yen. At lunch-
service time, restaurants with high rank
also offer less-expensive dishes, so that we
can enjoy them at relatively low cost. The
restaurant in Chuo University, where the
Symposium will take place, also offers
lunches around 500 Yen.
For dinners there is no special "service
time". We can usually have dinners for 800
Dishes with the above prices can be
found without any special effort. For those
who want to save more, we have many
shops for fast food such as hamburgers and
fried chicken, coin machines for canned
coffee, and supermarkets.
As for local transportation, subway
lines are in general most convenient. The
central part of Tokyo is covered with a
dense, and a little complicated, network of
subway lines, so that we can move from
one place to another only by subways.
There are several busy centers in central
Tokyo; for example Akasaka, Shinjuku,
Shibuya, Ikebukuro, and Ueno. These
centers are interconnected by subway lines,
and we can move between them for 120 or
140 Yen. JR line and many other private
lines connect the central part of Tokyo and
the suburbs, and an hour trip from central
Tokyo costs about 500 1,000 Yen. In
general the JR line is more expensive than
other private lines.
The daily rate of hotels reserved for the
meeting range from 5,900 to 20,000 Yen.
For details refer to the 2nd announcement
for the 13th International Symposium on
Mathematical Programming, which will be
sent to you if you contact:
Organizing Committee for the 13th
International Symposium on Mathe-
c/ The Operations Research Society of
2-4-16 Yayoi, Bunkyo-ku
(Editor's Note: On January 14, 1988 the
exchange rate was 126.2 Yen per US dollar.)
- - --- -- -- ----
0 P TI M A number twenty-three
Ca f P apr
Optimization Days 1
May 2,3,4, 1988
Abstract, including authors'
affiliations, should be limited to i
8'/, x 11 and submitted no later
January 31, 1988. Contact: Profe
Jacques A. Ferland, Dept. I.R.O.,
de Montr6al, C.P. 6128, Succ A, I
(Qu6bec) H3C 3J7, Canada.
Conference on Matri:
The Johns Hopkins U
sity, Baltimore, Mary
June 20-24, 1988
SThe principal lecturer, Robel
.'np-on, will give 10 lectures
qualities for eigenvalues, singular
and invariant factors with applic
control theory and functional an
Apply now, but not later than A
to ensure consideration for on-ca
housing and financial support.
Roger Horn, Department of Matl
Sciences, The Johns Hopkins Un
Baltimore, Maryland 21218.
Tenth Symposium on
t IT .1 1 .-_ --
ming with Data
A Tenth Symposium on Mathematical
Programming with Data Perturbations will
be held at The George Washington
Universitys Marvin Center on May 26-27,
1988. This symposium is designed to bring
ether practitioners who use mathemati-
a I programming optimization models and
deal with questions of sensitivity analysis,
with researchers who are developing
[988 techniques applicable to these problems.
CONTRIBUTED papers in mathematical
programming are solicited in the following
(1) Sensitivity and stability analysis results
and their applications.
(2) Solution methods for problems
names and involving implicitly defined problem
3oe page functions.
than (3) Solution methods for problems
ssor involving deterministic or stochastic
University parameter changes.
Montr6al (4) Solution approximation techniques and
-J. Ferland CLINICAL" presentations that describe
problems in sensitivity or stability analysis
C encountered in applications are also
ABSTRACTS of papers intended for
Jniver- presentation at the Symposium should be
land sent in triplicate to Professor Anthony V.
Fiacco. Abstracts should provide a good
technical summary of key results, avoid the
use of mathematical symbols and refer-
rt C. ences, not exceed 500 words, and include a
on ine- title and the name and full mailing address
tr values, of each author. The deadline for submis-
ations to sion of abstracts is March 11, 1988.
alysis. APPROXIMATELY 30 minutes will be
pril 8, 1988, allocated for the presentation of each paper.
impus A blackboard and overhead projector will
Contact: be available.
hematical Sponsored by: the Department of
diversity, Operations Research and the Institute for
Management Science and Engineering,
School of Engineering and Applied Science,
The George Washington University,
[ Washington, DC 20052. Phone: (202) 994-
International Meeting on
September 28th-30th, 1988-
The meeting will be held at Verona
University. The invited speakers are: D.
Bini (Italia), C. Canuto (Italia), M. Cosnard
(Francia), L. C. W. Dixon (Inghilterra), J.
Dongarra (U.S.A.), I. Duff (Inghilterra), D. J.
Evans (Inghilterra), I. Galligani (Italia), L.
Grandinetti (Italia), D. Parkinson (Inghil-
terra), A. Quarteroni (Italia), G. Radicati
(Italia), R. Rossi (Italia), U. Schendel
(Germania), F. Sloboda (Cecoslovacchia), F.
The main topics of the conference are:
Linear Algebra, Mathematical Program-
ming and Optimization, and Differential
The Organizing Committee is the
following: L. C. W. Dixon, I. Galligani, L.
Grandinetti, C. Nodari Sutti, D. Parkinson,
G. Radicati, R. Rossi.
Information requests and application
forms are to be addressed to:
c/o Istituto di Matematica
Via dell'Artigliere, 19
37129 Verona (ITALY)
Tel. + 39 45 8098216
A Call for Papers
The Canadian Journal of Administra-
tive Sciences is a multidisciplinary journal
published by the Administrative Sciences
Association of Canada. This journal is
regarded as one of the best of its kind
among the multidisciplinary journals in
The editor, Professor Jean-Charles
Chebat, welcomes research articles on
Management Science. Please forward your
papers addressed as follows:
Professor Jean-Charles Chebat
CANADIAN JOURNAL OF ADMINIS-
P. O. Box 8888, Station A
The area editor in Management Science
is Dr. Carl-Louis Sandblom, Technical
University of Nova Scotia.
There are no specific topics. Papers are
evaluated on an anonymous basis. The rate
of acceptance is approximately 1 out of 5.
~ I- ~---~-~ ~ ~~--------------`--~
0 P TI M A number twentV-three
*- Technical Reports & Working Papers
School of Operations Research
and Industrial Engineering
Ithaca, NY 14853
M. Phelan, "Some Problems of Estimation from Poisson Type
Counting Processes," TR 718.
Y. Herer, "Buffer Placement in Sequential Production Lines: Still
Further Studies," TR 719.
M. Todd, "Exploiting Special Structure in Karmarkar's Linear
Programming Algorithm," TR 720.
M. Todd, "Improved Bounds and Containing Ellipsoids in
Karmarkar's Linear Programming Algorithm," TR 721.
D. Heath and R. Kertz, "Leaving An Interval in Limited Playing
Time," TR 722.
D. Heath and W. Sudderth, "Coherent Inference from Improper
Priors and from Finitely Additive Priors," TR 723.
M. Hartmann, "A Note on the Tail Distribution of the Accumulated
Value of an Additive Functional of an Irreducible Markov Chain,"
J. Mitchell and M. Todd, "On the Relationship Between the Search
Directions in the Affine and Projective Variants of Karmarkar's Linear
Programming Algorithm," TR 725.
R. Bland and B. Dietrich, "A Unified Interpretation of Several
Combinatorial Dualities," TR 726.
G.A.P. Kindervater and J.K. Lenstra, "Parallel Computing in
Combinatorial Optimization," TR 727.
R. Bechhofer and C. Dunnett, "Subset Selection for Normal Means
in Multi-Factor Experiments," TR 728.
H. Kaspi, "Random Time Changes for Processes with Random Birth
and Death," TR 729.
R. Bland and D. Shallcross, "Large Traveling Salesman Problems
Arising from Experiments in X-Ray Crystallography: A Preliminary
Report on Computation," TR 730.
D. Joneja, "Multi-echelon and Joint Replenishment Production and
Distribution Systems with Nonstationary Demands," TR 731.
R. Bland and D. Jensen, "Weakly Oriented Matroids," TR 732.
R. Bland and D. Cho, "Balancing Configurations in R' by
Reflection of Points," TR 733.
P. Jackson, C. Jones and J. Muckstadt, "A Framework for a
Computer Aided Logistics System," TR 734.
P. Jackson, C. Jones and J. Muckstadt, "The COSMOS Sched-
uler," TR 735.
J. F. Claver and P.L. Jackson, "Lot Sizing in Cyclic Scheduling,"
L. Trotter and D. Crystal, "On Abstract Integral Dependence,"
M. Phelan and N.U. Prabhu, "Estimation from an Infinite Server
Queueing System with Two Demands," TR 738.
J.S. Mitchell, "Shortest Rectilinear Paths Among Obstacles,"
E. Yucesan and L. Schruben, "Transaction Tagging in Highly
Congested Simulation Models," TR 740.
J. Mitchell and M.J. Todd, "Two Variants of Karmarkar's Linear
Programming Algorithm for Problems with Some Unrestricted Vari-
ables," TR 741.
R. Bland, "Combinatorial Perspectives on Linear Programming,
Part I: Linear Programming Duality and Minty's Lemma," TR 742.
L. Trotter and S. Tipnis, "A Generality of Robacker's Theorem,"
R. Bechhofer and D. Goldsman, "Truncation of the Bechhofer-
Kiefer-Sobel Sequential Procedure for Selecting the Normal Population
which has the Largest Mean," TR 744.
R. Bechhofer, C. Dunnett and A. Tamhane, "Optimal Allocation
of Observations in Subset Selection and Multiple Comparisons with a
Control, and Associated Tables (with Application to Drug Screening),"
E. Arkin and J.S. Mitchell, "An Optimal Visibility Algorithm for a
Simple Polygon with Star-Shaped Holes," TR 746.
B. Fox, "Gradient Computation for Transient Markov Chains,"
J. Renegar, "On the Worst Case Arithmetic Complexity of Approxi-
mating Zeros of Systems of Polynomials," TR 748.
N.U. Prabhu, "A Theory of Semi-regenerative Phenomena,"
R. Freund, "An Analogue of Karmarkar's Algorithm for In.Pliili
Constrained Linear Programs, with a "New" Class of Projective Tran, .
nations for Centering a Polytope," TR 750.
R.O. Roundy, "Computing Nested Reorder Intervals for Multi-Item
Distribution Systems," TR 751.
P.L. Jackson, C. Jones and J. Muckstadt, "COSMOS-The Cornell
Simulator of Manufacturing Operations," TR 752.
B. Fox, "Complexity of Gradient Estimation for Transient Markov
Chains," TR 753.
R. Barton, "Testing Strategies for Simulation Optimization,"
B. Fox, "Computing the Gradient of Expected Reward Up to
Absorption," TR 755.
University of Maryland at College Park
College of Business and Management
Management Science and Statistics
College Park, MD
B. Golden and E. Wasil, "Ranking Outstanding Sports Records,"
M. Ball and M. Magazine, "Sequencing of Insertions in Printed
Circuit Board Assembly," WP87-002.
L. Bodin and D. Salamone, "The Development of a Microcomputer
Based System for Vehicle Routing and Its Use for Solving Spatial and .
Temporal Problems," WP87-003.
B. Golden, Q. Wang and L. Liu, "A Multi-Faceted Heuristic for
the Orienteering Problem," WP87-006.
0 PT IM Anumber twenty-three
0 P T I M A number twenty-three
C. Skiscim and B. Golden, "Solving k-Shortest and Constrained
Shortest Path Problems Efficiently," WP87-009.
A. Assad, "History of Operations Research: Some Anticipatory
R. Dahl, K. Keating and L. Levy, "WHAMII: An Enumeration
and Insertion Procedure with Binomial Bounds for the Stochastic Time-
Constrained Traveling Salesman Problem," WP87-012.
R. Vohra and B. Golden, "Finding the Most Vital Arcs in a
S. Gass, "Operations Research Supporting Decisions Around the
i '...!,,." WP87-017.
M. Ball and H. Benoit-Thompson, "A Lagrangian Relaxation
Based Heuristic for the Urban Transit Crew Scheduling Problem,"
A. Assad and C. Skiscim, "The Maximal Covering Location
Problem and Dual Ascent," WP87-020.
M. Ball, L. Liu and W. Pulleyblank, "Two Terminal Steiner Tree
D. Casco, B. Golden and E. Wasil, "Vehicle Routing with
M. Ball, U. Derigs and C. Hilbrand, "Matching Problems with
Generalized Upper Bound Side Constraints," WP87-026.
N. Collins, R. Eglese and B. Golden, "Simulated Annealing An
Annotated Bibliography," WP87-028.
,'slemn.i Optimization Laboratory
Department of Operations Research
Ian ford University
Stanford, CA 94305-4022
B. C. Eaves and U. G. Rothblum, "Invariant Polynomial Curves of
Piecewise Linear Maps," SOL 86-13.
B. C. Eaves and U. G. Rothblum, "A Flexible Manufacturing and
Operator Scheduling Model Solved by Deconvexification Over Time,"
B. C. Eaves, "Thoughts on Computing Market Equilibrium with
SLCP," SOL 86-15.
G. B. Dantzig, P. H. McAllister and J. C. Stone, "Using the
PILOT Model to Study the Effects of Technical Change," SOL 86-16.
G. B. Dantzig, "Planning Under Uncertainty Using Parallel
Computing," SOL 87-1.
G. B. Dantzig and J. A. Tomlin, "E.M.L. BEALE, FRS: Friend and
Colleague," SOL 87-2.
K. A. E. Ercikan and F. S. Hillier, "Heuristic Procedures for 0-1
Integer Programming," SOL 87-3.
I. J. Lustig, "The Equivalence of Dantzig's Self-Dual Parametric
Algorithm for Linear Programs to Lemke's Algorithm for Linear Comple-
mentarity Problems Applied to Linear Programs," SOL 87-4.
G. B. Dantzig, "Origins of the Simplex Method," SOL 87-5.
K. Zikan and R. W. Cottle, "The Box Method for Linear Program-
ming: Part I-Basic Theory," SOL 87-6.
R. Entriken, "Using MINOS as a Subroutine for Decomposition,"
I. J. Lustig, "Comparisons of Composite Simplex Algorithms,"
K. Zikan and R. W. Cottle, "The Box Method for Linear Program-
ming: Part II-Treatment of Problems in Standard Form with Explicitly
Bounded Variables," SOL 87-9.
C. Fraley, "Computational Behavior of Gauss-Newton Methods,"
I. J. Lustig, "An Analysis of an Available Set of Linear Program-
ming Test Problems," SOL 87-11.
P. E. Gill, W. Murray, M. A. Saunders and M. H. Wright, "A
Schur-Complement Method for Sparse Quadratic Programming,"
Y. Yinyu, "Eliminating Columns in the Simplex Method for Linear
Programming," SOL 87-14.
SJournals & Studies
Vol.40, No. 2
J. Reneger, "Rudiments of an Average Case Complexity Theory for
Piecewise-Linear Path Following Algorithms."
F. Glover and D. Klingman, "Layering Strategies for Creating
Exploitable Structure in Linear and Integer Programs."
D. Goldfarb and S. Mehrotra, "A Relaxed Version of Karmarkar's
A.J. Hoffman, "On Greedy Algorithms for Series Parallel Graphs."
G.G.L. Meyer, "Convergence of Relaxation Algorithms by Averag-
Vol. 40, No. 3
B. Fleischmann, "A New Class of Cutting Planes for the Symmetric
Travelling Salesman Problem."
R. H. Byrd, R. B. Schnable and G. A. Shultz, "Approximate
Solution of the Trust Region Problem by Minimization over Two-
R. H. F. Jackson and G. P. McCormick, "Second-Order Sensitivity
Analysis in Factorable Programming: Theory and Applications."
A.V. Fiacco and J. Kyparisis, "Computable Bounds on Parametric
Solutions of Convex Problems."
D. Goldfarb and S. Mehrotra, "Relaxed Variants of Karmarkar's
Algorithm for Linear Programs with Unknown Optimal Objective
Y. Pochet and L. A. Wolsey, "Lot-Size Models with Backlogging:
Strong Reformulations and Cutting Planes."
--~--~ ~ ---- ~- ~
PAE8 TIM ume tetytre EBURY18
By L. Lovasz and M. D. Plummer
Annals of Discrete Mathematics 29
North-Holland, Amsterdam, 1986
Today, combinatorial optimization is considered to be one of the
"hot fields" within mathematics. Publication rates almost explode,
keeping the scientists of that field in a permanent hurry, which quite
often results in a considerable decrease of quality and/or clarity of
presentation in publications. From this point of view, Matching Theory
by L. Lovasz and M. D. Plummer appears to be written outside of time.
Everything is developed with splendid clarity, organized in a master-
ful way and written in an excellent style, sometimes exciting, some-
times humorous, always maintaining lively dialogue with the reader.
This beautiful book (in fact the first that has ever been written on
Matching Theory) provides a comprehensive treatment of the subject,
leading up to the frontiers of current research. (Unfortunately,
however, the recent results of the first author about matching lattices
were obtained shortly after completion of the volume and hence do
not appear in this book, while the main tools for obtaining them, i.e.
ear- and brick decomposition, can already be found there).
The introduction provides an interesting accountof thehistoryof
matching theory, starting with the "Founders of the discipline", J.
Petersen and D. K6nig. The book itself is divided into 12 chapters: 1-
Bipartite matching, 2-Flow theory, 3-Size and structure of maximum
matching, 4-and 5-Bipartite (general) graphs with perfect matching,
6-Some graph theoretical problems related to matching, 7-Matching
and linear programming, 8-Determinants and matching, 9-Matching
algorithms, 10-The f-factor problem, 11-Matroid matching, and 12-
Vertex packing and covering.
The authors concentrate on structural problems and polynomial
solvability. Consequently, nothing is said about implementational
aspects or parallel complexity of the matching problem (although
probabilistic methods are sketched in chapter 8. On the other hand,
several cross connections to related subjects are at least outlined
whenever this seems to be worthwhile to do. So, in fact, one can learn
somewhat more than just matching theory from this book. (For
example, you will probably be surprised that matching may be used
for proving the existence of Haar measure on compact topological
groups, to mention one of the more "exotic" applications). Whenever
a general idea originating in a different field is used somewhere, this
is explained afterwards. For this purpose, a set of "boxes" has been
included, containing some relevant material from other fields (e.g.
complexity theory, matroid theory etc.) in a condensed manner.
These may easily be skipped by the reader with more background
information a good idea!
Summarizing, this book provides an excellent introduction to
Matching Theory. It may well serve as a textbook for students, but it
contains some interesting material for experts as well. The list of
references given at the end contains about 500 entries. However, as
has been mentioned in the introduction of this review, this is, of
course, far from being complete and a more extensive bibliography
has been announced by the authors.
Algorithmic Graph Theory
By Alan Gibbons
Cambridge University Press, Cambridge, 1985
It is my feeling that there will be never enough books on graph
theory and their applications. Indeed, the whole field is diverse and
does not form a "theory" in the classical sense. Moreover, due to the
steady flow of new ideas from other fields and praxis, graph theory is
a rapidly changing subject.
This book centers on the algorithmic aspects of graphs. This is not
a novel approach as there are several books with more or less the same
orientation. Yet this book brings something new: it presents a very
balanced introductory text which stresses the algorithmic and com-
The organization of the material is mainly traditional as indicated
by headings of Chapters:
1. Introducing graphs and algorithmic complexity; 2. Spanning trees,
branchings and connectivity; 3. Planar graphs; 4. Network and flows;
5. Matchings; 6. Eulerian and Hamiltonian tours; 7. Colouring graphs;
8. Graph problems and intractability; Appendix: On linear program-
However the contents are less traditional and let us list several
particulars: Chapter 1 contains a nice discussion of the depth first
searching of graphs; Chapter 2 covers not only spanning trees but also
optimum branchings; Chapter 4 on network flows covers the mini-
mum -cost flow algorithm; Chapter 5 has maximum weighted match'
ing in general graphs; Chapter 6 discusses the Chinese postman
problem; Chapter 8 contains several NP complete results including
the recent Holyer result on factorization of cubical graphs.
0 P T IM A number twenty-three
1'AG 9 0 P ubrtet-he ERAY18
Nearly each chapter contains a proof of a deeper result sometimes
unexpected for a text of this size (to name just a few examples:
Kuratowski's theorem, Cook's theorem, and non-bipartite weighted
I think that the text is a useful addition to the existing literature.
It is worthwhile as an introductory text for undergraduate and early
graduate students of both mathematics and computer science.
The Travelling Salesman Problem
By E. L. Lawler et. al.
Wiley, Chichester, 1985
Combinatorial optimization is coming of age, in fact
1) there has been an impressive growth in its body of knowledge
and applications since the first pioneering results became known
some thirty years ago; also, the field is attracting more and more
researchers: the handful of friends that started it has meanwhile
become a sizeable community spread over the whole world;
2) it is providing a common ground for research in all kinds of
disciplines such as physics, economics, engineering, computer sci-
3) there is an increasing unification and standardization in its
terminology and notation, enhanced by the ever growing number of
high standard specialized journals;
4) last but not least, note that of late, several outstanding textbooks
(the above being an excellent case in point) are giving a wider public
access to combinatorial optimization.
The book on the TSP, edited by Lawler, Lenstra, Rinnooy Kan and
Shmoys is a delight to read. It claims to be a guided tour of combina-
torial optimization, but it really is much more than just that. It is a full
grown textbook for the beginner as well as a source of inspiration and
reference for the more advanced reader. Not only does it reflect the
coming of age of combinatorial optimization, but also the vitality of
this discipline with its wealth of challenging old and new open
One of the achievements of the book, only partially explained by
'em 3) above, is its high degree of "unite de doctrine". Indeed, the
F Pok is not a mere collection of surveys, but a well structured,
balanced and largely self-contained treatise of some of the major
results in combinatorial optimization, sewn together by the travelling
Yet, the book certainly does not mark the end of the travelling
salesman's journey, insofar as more and more of its hidden facets are
being discovered by the day. Also, the state of the art regarding the
exact solution as well as heuristics has not stood still since the
publishing of the book. See, e.g. Padberg/Rinaldi's Branch & Cut
code solving a TSP problem with 2392 cities to optimality, or the not
quite as spectacular, nevertheless efficient, simulated annealing
As will be seen in the list of contents given below, the contribu-
tors to the book are some of the most respected and articulate research-
ers in the field. The editors may be congratulated for their good taste.
Moreover, the book has quite certainly profited from the action of the
editors' invisible warning finger. Its chapters overlap just enough to
add to its coherence, without making it tedious by unnecessary
repetitions. Also, and quite unusually, the reader is not unduly
irritated or even frustrated by the authors inviting her to fill in
"obvious" details that were left out.
1. History, by A. J. Hoffman and P. Wolfe: A sketch of the history of
the TSP from Euler, Kirkman and Hamilton to present times. The TSP
is convincingly argued to be a deep problem, by noting e.g. that
several methods that shaped mathematical programming, were actu-
ally devised in order to tackle it.
2. Motivation and modeling, by R. S. Garfinkel: Or, how to cut
wallpaper such that patterns match and waste is minimized, as well
as many other problems modeled by the TSP.
3. Computational complexity, by D. S.Johnson and C. H. Papadimi-
triou: A remarkable introduction to the general topic of complexity
theory and the analysis of algorithms, going well beyond the TSP,
introducing P, NP and NP-completeness proof techniques.
4. Well-solved special cases, by P. C. Gilmore, E. L. Lawler and D. B.
Shmoys: The bright side of TSP: nontrivial special cases that can be
solved efficiently, alas a slowly growing class.
5. Performance guarantees for heuristics,by D. S. Johnson and C. H.
Papadimitriou: A continuation of chapter 3, with the complexity and
worst case performance analysis of several heuristics based on vari-
ous relaxations of the TSP.
6. Probabilistic analysis of heuristics, by R. M. Karp and J. M. Steele:
"The length of the shortest tour connecting a cloud of random points
on the unit square tends to behave a.s. like a deterministic function
proportional to the square root of the number of points, as this number
I ~ --- -~
0 P TI M A number twenty-three
PAGE 10 0 P T I M A number twenty-three FEBRUARY 1988
J BOOK REVIEWS
gets large enough". This classic result by Beardwood et al. leads to an
efficient heuristic for the Euclidean TSP. A result of a similar flavour
by Walkup is used to analyze a patching algorithm for the asymmetric
TSP with uniformly i.i.d. data.
7. Empirical analysis of heuristics, by B. L. Golden and W. R.
Stewart: A description and a statistical performance analysis of
various heuristics. Estimating a confidence interval for the value of
the optimal solution of the TSP based on outputs from heuristics.
8. Polyhedral theory, by M. Gr6tschel and M. W. Padberg: A lucid
and self contained introduction to polyhedral theory in general and to
what is known thanks to the authors concerning the TSP in particular.
9. Polyhedral computations, by M. W. Padberg and M. Gr6tschel: A
review of the equivalence of a "good characterization" of the convex
hull of a discrete optimization problem and the existence of an
efficient algorithm to solve it, with a proof based on the ellipsoid
method. How to efficiently identify certain violated constraints
(facets) of the convex hull of the TSP. Finally a report on remarkable
computational experiences using these inequalities in conjunction
with branch and bound codes.
10. Branch and bound methods, by E. Balas and P. Toth: A general
introduction to Branch and Bound, a method whose co-fatherhood
could rightly be claimed by the first author. A review of the various
possible relaxations in conjunction with the TSP. Performance of their
and others' state-of-the-art computer codes.
11. Hamiltonian cycles, by V. Chvatal: "This is a biased exposition,
meant to introduce the reader to Hamiltonian graphs rather than to
provide an extensive survey of their theory" (quotation from the
conclusion of this chapter and an understatement). A very stimulat-
ing essay, showing that recognizing Hamiltonian graphs can be
12. Vehicle routing, by N. Chistofides: A review of models and
heuristics for this important extension of the TSP.
Bibliography: Comprising over 400 items ranging from Adrabinski
In summary: Buy the book, you won't regret it!
-Th. M. Liebling
STORM: Quantitative Modelling for
By H. Emmons, A. D. Flowers, K. Marthn,
C. M. Khot
What's Best!: Decision Support Software
By S. L. Savage
Holden Day, Oakland, 1986
In recent years personal computers have undergone a rapid
improvement of storage and speed. This development has prompted
attempts to write Operations Research software for this class I.
computers. These programs can serve two purposes. First it makes
Operations Research available to smaller firms and/or quick deci-
sions. For instance a controller might set up a small linear program to
analyse a certain situation (this is practised e.g. in a special course at
the Controller Akademie in Ganting, W. Germany). Secondly it might
be advantageous to use Operations Research software in an introduc-
tory (or advanced) university course on the topic. This could make the
tedious "m= 3-n =4 -L.P- by-hand-solving" obsolete. Here we review
two of these products. "Storm" seems to be better suited for the latter
purpose. It is a program package containing all the Operations
Research models we learn about in a beginner's Operations Research
course. For a practitioner it might be a little bit too much and a little
bit too academic. The topics which can be treated are:
1) Linear Programming (including sensitivity analysis and par-
2) The Assignment Problem (solved by the Hungarian method)
3) The Transportation Problem (solved by a network simplex
method; a capacitated version and one with bounds on supply and
demand are included)
4) Queueing Analysis ((M/M/C), (M/D/C), (M/G/C), (M/M/
5) Project Management (CPM/PERT)
6) Inventory Management (based on the EOQ-model and its exten-
0 PT IM Anumber twenty-three
PAG 10PTIMAnme wet-he ERAY18
7) Facility Layout (this is certainly not necessary for quick decision
8) Assembly Line Balancing
9) Investment Analysis
11) Production Scheduling
12) Material Requirements Planning
13) Regressions Analysis
Even though this list reads like the contents of an Operations
Research textbook, the authors have not been able to agree on a single
reference for a single chapter in the manual (e.g. for two different
queueing models they quote two different sources).
While "Storm" is mainly a collection of Operations Research
programs which run on a PC, "What's best!" is especially designed to
be used on and with a PC. "What's best!" can solely solve linear
programs. However it seems to me to be made mostly for use in
business decisions. Its connection with the Lotus 1-2-3 spreadsheet
concept seems to enhance creating programs out of available data.
"What's best!" can be a powerful tool.
Overall it can be hoped that software systems like the two
reviewed above will increase the importance of Operations Research
concepts for middle sized firms.
Application for Membership
THE MATHEMATICAL PROGRAMMING
do International Statistical Institute
428 Prinses Beatrixlaan
2270 AZ Voorburg
Cheques or money orders should be made payable to
THE MATHEMATICAL PROGRAMMING SOCIETY, INC.
in one of the currencies listed below.
Dues for 1988, covering subscription to volumes 40-42 of
Mathematical Programming, are Dfl.115.00 (or $55.00 or
DM105.00 or 35.00 or FF325.00 or Sw.Fr.85.00).
'cudent applications: Dues are one-half the above rates.
Subscription to the Studies is at full rates. Have a faculty
member verify your student status and send application
with dues to above address.
I wish to enroll as a member of the Society. My subscrip-
tion is for my personal use and not for the benefit of any
library or institution. I enclose payment as follows:
Dues for 1988
Name (please print)
Mailing Address (please print)
Faculty verifying status
I_~ ~ I ----~
OPTIMA number twenty-three
OP T I M A number twenty-three
The 1986 Lanchester Prize was awarded to Alexan-
der Schrijver (Tilberg) and Peter Whittle (Cam-
bridge) at the ORSA/TIMS meeting in St. Louis.
Schrijver received the prize for his research
monograph Theory of Linear and Integer
Programming and Whittle for the monograph Sys-
tems in Stochastic Equilibrium. Both monographs
were published by Wiley in 1986...At the same
meeting George Nemhauser (Georgia Tech) gave
the plenary address and was given an honorary
membership in Omega Rho...Jean-Louis Goffin
(McGill) is on sabbatical at CORE, University
Catholique de Louvain in Belgium...James Ho
(Tennessee) is visiting the OR Department at Stan-
ford during Spring, 1988...Siriphong ("Toi")
Lawphongpanich has joined the OR Department at
the Naval Postgraduate School as an assistant pro-
Deadline for the next O P T I M A is June 1, 1988.
I i-- --- - -- - --- --
P T I M A
MATHEMATCAL PROGRAMMING SOCIETY
303 Weil Hall
College of Engineering
University of Florida
Gainesville, Florida 32611
FIRST CLASS MAIL
ill-L-IXli^Ciil~ilX-~~-3 ~- r=I
Books for review should be
sent to the Book Review Editor,
Prof. Dr. Achim Bachem,
Mathematiches Institute der
Universitdit zu K61n,
Weyertal 86-90, D-5000 K61n,
Journal contents are subject
to change by the publisher.
Donald W. Hearn, Editor
Achim Bachem, Associate Editor
Published by the Mathematical
Programming Society and
Publication Services of the
College of Engineering,
University of Florida.