PTI MA
MATHEMATICAL PROGRAMMING SOCIETY NEWSLETTER
No29
March 1990
Amsterdam 1991
A MSTERDAM, the capital of the Netherlands, will
host the 14th International Symposium on
Mathematical Programming from August 5 to
9,1991.This is the triennialmeetingoftheMathe
matica l Programming Society, andwe cordially
invite you to participate in the meeting and to contribute to
making it a success.
Amsterdam is known all over the world for a variety of reasons.
Its historic center covers a wide area of, mainly 17th Century,
buildings along "grachten" (canals). The cultural attractions in
clude theRijksmuseum, with several Rembrandts (Nachtwacht),
Vermeers, Frans Ha Is, etc.; the Vincent van Gogh Museum, with
a huge collection of van Goghs; the StedelijkMuseum,formodern
art; the Royal Concertgebouw Orchestra, and the National Bal
let. Moreover, we mention the liberal atmosphere, yielding a rich
supply of pleasures of various kinds.
The Symposium will takeplace in the buildings of the University
of Amsterdam in the city center, with several hotels and restau
rants of diverse categories nearby. The meeting offers invited and
contributed talks in parallel sessions, and we call for papers on
all theoretical, computational, andpractical aspects four field.
We have chosen a particularly late deadline for the submission
of papers (June 1, 1991) so as to encourage the presentation of
very recent results.
Last October we sent out the First Announcement to a large
number of people, including all members of the Society. It con
tains some more information and is reprinted in this issue of
OPTIMA. The Second Announcement will appear this Fall and
will be sent to all those who have returned the Preregistration
Form in the First Announcement.
We look forward to seeing you in Amsterdam and to offering
you a pleasant and fruitful symposium.
JAN KAREL LENSTRA
ALEXANDER RINNOOY KAN
ALEXANDER SCHRIJVER
Council Meets in
the Black Forest
Bernhard Korte and Klaus Ritter
were the organizers of the VIth
Biennial Oberwolfach Confer
ence on Mathematical Program
ming. Between 60 and 70 mathe
matical programmers met from
January 7 to January 13 at the
"Mathematisches Forschungsin
stitut" in Oberwolfach. They
attended more than 50 presenta
tions on recent research; they
enjoyed the informal atmos
phere and the scenic surround
ings; and many participated in
the traditional hike on Wednes
day afternoon.
CONTINUES, PAGE FOUR
OPTIMA
NUMBER 29
CONFERENCE NOTES 3
TECHNICAL REPORTS &
WORKING PAPERS 56
BOOK REVIEWS 710
JOURNALS 11
GALLIMAUFRY 12
1 I I ~ ~  ~ ' 
e w n ne M
14th International Symposium on Mathematical Programming
Amsterdam, The Netherlands, August 59, 1991
First Announcement
The International Symposium on Mathematical Programming is the triennial scientific meeting of the Mathematical Programming Society.
The 14th Symposium will be held at the University of Amsterdam, August 59, 1991. It is jointly organized by the University of Amsterdam,
the Centre for Mathematics and Computer Science (CWI) in Amsterdam, Eindhoven University of Technology, and the Erasmus University
in Rotterdam.
Chairmen: J. K. Lenstra, A. H. G. Rinnooy Kan, A. Schrijver. International Program Committee: A. Auslender, M. Avriel, E. Balas, M. L. Balinski,
C. Berge, R.E. Bixby, V. ChvAtal, A. R. Conn, R.W. Cottle, G.B. Dantzig, J. E. Dennis,Jr., L.C.W. Dixon, A. M. Geoffrion, F. Giannessi, J.L. Goffin,
D. Goldfarb, E. G. Golshtein, R. L. Graham, M. Gr6tschel, P. L. Hammer, M. Held, A.J. Hoffman, K. L. Hoffman, T. Ibaraki, M. Iri, E. L.Johnson, P. Kall,
R. Kannan, R. M. Karp, A. V. Karzanov, M. L. Kelmanson, V. Klee, K. 0. Kortanek, B. Korte, J. Krarup, H. W. Kuhn, E. L. Lawler, C. Lemar6chal,
F. A. Lootsma, L. LovAsz, O. L. Mangasarian, G. P. McCormick, N. Megiddo, G. Mitra, B. Mond, G. L. Nemhauser, W. Oettli, M. W. Padberg, B. T. Polyak,
M.J.D. Powell, A. Pr6kopa, W. R. Pulleyblank, M.R. Rao, K. Ritter, S.M. Robinson, R. T. Rockafellar, J.B. Rosen, H.E. Scarf, R. B. Schnabel,
P. D. Seymour, J. Stoer, E. Tardos, M.J. Todd, A. W. Tucker, H. Tuy, A. F. Veinott,Jr., R.J.B. Wets, A. P. Wierzbicki, P. Wolfe, L. A. Wolsey, M.Y. Yue.
Organizing Committee: O.J. Boxma, A. M. H. Gerards, J. L. dejong, G. A. P. Kindervater, M. Labbe, B.J. Lageweg, G. de Leve, M. W. P. Savelsbergh,
A.J.J. Talman, H. C. Tijms, L. N. van Wassenhove. Advisory Committee: M. Grotschel (chairman), H. Konno, R. B. Schnabel, M.J. Todd.
Call for papers. Papers on all theoretical, computational and
practical aspects of mathematical programming are welcome.
The presentation of very recent results is encouraged. For this rea
son, a particularly late deadline for the submission of titles and
abstracts has been set.
Dates and deadlines
September, 1990: Second Announcement
April 1, 1991: Deadline for early registration
June 1, 1991: Deadline for submission of titles and abstracts
August 59, 1991: The Symposium
Topics. Sessions on the following topics are organized. Sugges
tions for further areas to be included are welcome.
 Linear, integer, mixedinteger programming
 Interiorpoint and pathfollowing algorithms
 Nonlinear, nonconvex, nondifferential, global optimization
 Complementarity and fixed point theory
 Dynamic and stochastic programming, optimal control
 Game theory and multicriterion optimization
 Combinatorial optimization, graphs and networks, matroids
 Computational complexity
 Approximative methods, heuristics
 Computational geometry, VLSIdesign
 Implementation and evaluation of algorithms and software
 Largescale mathematical programming
 Parallel computing in mathematical computing
 Expert, interactive and decision support systems
 Mathematical programming on personal computers
 Teaching in mathematical programming
 Applications of mathematical programming in industry,
government, economics, management, finance, transportation,
engineering, energy, environment, agriculture, sciences and
humanities
Structure of the meeting. The meeting will offer invited and con
tributed lectures in parallel sessions. In addition, computer
demonstrations and survey lectures highlighting developments of
current interest are planned. During the plenary opening session,
the George B. Dantzig Prize (for original research with a major
impact on mathematical programming), the Fulkerson Prizes (for
outstanding papers in discrete mathematics), the OrchardHays
Prize (for excellence in computational mathematical program
ming), and the A.W. Tucker Prize (for an outstanding paper by a
student) will be awarded. The program also contains a reception
and a banquet.
Site. The Symposium will take place in the Oudemanhuispoort
building of the University of Amsterdam, located in the historic
centre of Amsterdam, close to many attractions of various kinds.
Amsterdam is easily reachable by all means of public transporta
tion and has direct air connections with many cities all over the
world.
Preregistration. The Second Announcement, which will appear
in September 1990, will be sent to all those who return the
Preregistration Form below. It will contain additional and more
detailed information about the program, registration fees, social
events, hotels, traveling, etc.
Mailing address
14th International Symposium on Mathematical Programming
Paulus Potterstraat 40
1071 DB Amsterdam
The Netherlands
Telephone: +3120752120
Telefax: +31206628136
Telex: 10761 omega.nl
Electronic mail: ismp@swivax.uucp, ismp@swi.psy.uva.nl
~ __s I ~I~ ~
MARCH 1990
PAGE 2
number twentynine
Pnm wtnA 1
First
Preregistration Form
Please keep me on the mailing list for further information about the 14th International Symposium on Mathematical Programming.
name: O I plan to attend the Symposium.
name: I intend to give a talk.
Mailing address:
Institution (if not in mailing address):
My talk may be on the following subject:
Please return this form to:
14th International Symposium on Mathematical Programming
Paulus Potterstraat 40
1071 DB Amsterdam
The Netherlands
Nato Advanced Study Institute
on Combinatorial Optimization
Bilkent University
Ankara, Turkey
July 1628, 1990
NATO is sponsoring an ASI, "New Fron
tiers in the Theory and Practice of Combina
torial Optimization: Applications in VLSI
Design," to be held at Bilkent University in
Ankara, Turkey, July 1628, 1990. The ASI is
also sponsored by the University of Florida
(S. Tufekci), Bilkent University (M. Akgul)
and the University of Kaiserslauten (H. W.
Hamacher). The objective of this institute is
to disseminate the stateoftheart knowl
edge on combinatorial optimization with a
focus on the applications in VLSI design.
This twoweek institute will be primarily in
the form of a workshop with lectures from
prominent, internationally renowned
scientists. It will be followed by an after
noon poster session where a limited number
of papers on the theory and practice of
combinatorial optimization will be
discussed. Papers might be submitted for
these poster sessions on network optimiza
tion, recent advances in linear program
ming, integer programming, traveling
salesman problem, parallel algorithms,
matroids, polyhedral combinatorics, and
application of combinatorial optimization
on manufacturing decision problems
including VLSI design. There is a limited
number of financial grants available for
participants from the NATO countries. To
be considered for a financial grant, appli
cants must provide background information
and a letter of recommendation from their
department head or from their dissertation
advisor no later than March 15, 1990. The
award notification will be mailed by May 5,
1990. Send applications and abstracts to DR.
SULEYMAN TUFEKCI, Associate Director,
Center for Optimization and Combinatorics
(COCO), Department of Industrial and
Systems Engineering, University of Florida,
Gainesville, Florida, 32611, USA.
Nordic MPS Meeting and
Formation of Geographical
Section
Some MPS members from the Nordic
countries will arrange a meeting on "Algo
rithms and Solution Procedures in Mathe
matical Programming" on August 25 and 26,
1990 in Copenhagen. The meeting is sup
ported in part by The Nordic Council of
Ministers and has two goals:
1. To increase contact between math program
mers in the Nordic countries;
2. To discuss the formation of a Nordic section
of MPS.
Applications for participation and abstracts
must be sent to Stein W. Wallace, Haugesund
Maritime College, SkAregaten 103, N5500
Haugesund, Norway, by April 30,1990. For
more information, please contact him at the
above address, or by phone (+47 4 721200) or
FAX (+47 4 715906). (email is not functioning
properly at this time.) It is possible to apply
for partial coverage of travel costs for those
who do not have other sources.
STEIN W. WALLACE
~irs~
   
MARCH 1990
PAGE 3
number twentynine
S4 nI
COUNCIL from page one:
A mong the attendants were all of the
members of the Council of the Mathematical
Programming Society: George Nemhauser
(Chairman), Michel Balinski (Past Chairman),
Les Trotter (Treasurer), Egon Balas, Bill Cun
ningham, Claude Lemarechal, and Alexander
Schrijver (Council MembersatLarge). It was
therefore decided to organize a rare event a
Council meeting in between the triennial in
ternational symposia. This meeting took place
on January 10 and was also attended by Bob
Meyer (Chairman of COAL), Clyde Monma
(Chairman of the Advisory Committee for the
1994 Symposium), Bill Pulleyblank (Editor of
Series B of the Journal), Mike Todd (Past Edi
tor of Series A of the Journal), Laurence
Wolsey (Chairman of the Publications Com
mittee), and Jan Karel Lenstra (Chairman of
the Executive Committee). I will summarize
the issues that were discussed below.
Meetings The 14th International Symposium
on Mathematical Programming will be held in
Amsterdam in 1991. You will find more about
it elsewhere in this issue. In the meantime, a
site for the 1994 symposium has to be selected.
A call for proposals has already appeared in
the previous issue of OPTIMA. Since then, the
Symposium Advisory Committee has sent
invitations to submit such a proposal to about
12 of our colleagues at American universities.
The Council suggested that our triennial
symposia might be held under cosponsor
ship of SIAM; the Chairman will explore this
idea.
The Society is a cosponsor of the Conference
on Integer Programming and Combinatorial
Optimization that is being organized by Ravi
Kannan and Bill Pulleyblank and will be held
in Waterloo at the end of May 1990. The Coun
cil felt that the Society should be catalytic in
this respect and encourage others who might
want to organize similar meetings.
Publications Due to an overflow of papers
that have been accepted for Series A of the
Journal, it is difficult to avoid a large backlog
for Series A while maintaining a regular pub
lication schedule for Series B. The Council
discussed several remedies. As a result, we
have entered negotiations with our publisher,
NorthHolland. There will most likely be a
substantial increase in the total annual vol
ume of Series A over the next few years.
Committee on Stochastic Programming The
Council approved of new membership of this
committee: John Birge, Michael Dempster,
Jitka Dupavcova (Secretary), Yuri Ermoliev,
Kurt Marti, Andras Prekopa, Yves Smeers,
Tomas Szantai, Roger Wets (Chair), and Wil
liam Ziemba.
Membership The Council invited the Chair
man to appoint a general Membership Com
mittee. It will have the task of advising the
Council on initiatives that could increase our
membership, e.g., by the creation of geo
graphical sections.
Egon Balas and Claude Lemarechal agreed to
serve on an ad hoc Committee for Special
Membership Arrangements. This committee
will advise the Council on special arrange
ments for members from countries with non
convertible currencies. At present, we have an
arrangement for nonpaying Hungarian
members, and we also received suggestions
for a Soviet membership from Professors
Golshtein and Levner. The Council prefers a
uniform policy to arrangements on a country
bycountry basis and feels that any arrange
ment of this kind should be financially reason
able for the Society and subject to periodic
(e.g., triennial) review.
Prizes The Council decided to change the
name of the OrchardHays Prize to the Beale
OrchardHays Prize. The membership of our
four prize committees has already been an
nounced in OPTIMA.
Administrativelssues The 1990 membership
list will contain information (telephone and
fax numbers, electronic addresses) that has
recently been collected. The Council decided
to give the members the option of paying their
dues by credit card, but decided against the
possibility of offering a threeyear member
ship at a reduced fee.
Algorithms and the Law The Council ex
pressed concern about the fact that the interior
point approach for solvingresourceallocation
problems has recently been patented in the
U.S. and that some of these patents might
misrepresent the history of our field. The
Chairman was urged to appoint a Committee
on Algorithms and the Law. This committee
will be asked to investigate the situation and
to advise the Council on possible courses of
action.
JAN KAREL LENSTRA
 II "~
MARCH 1990
PAGE 4
number twenty/nine
PAGE 5 number twentynine MARCH 1990
S
Technical
Reports &
Working
Papers
Georgia Institute of Technology
School of Industrial and Systems
Engineering
Atlanta, GA 30332
G.L. Nemhauser and R. Rushmeier, "Perform
ance of Parallel BranchandBound Algorithms
for the Set Covering Problem," J8902.
G.L. Nemhauser and G. Sigismondi, "A
Strong Cutting Plane/BranchandBound
Algorithm for Node Packing," J8908.
G.L. Nemhauser, G. Sigismondi and P.
Vance, "A Characterization of the Coefficients in
FacetDefining Lifted Cover Inequalities," J89
06.
G. Parker and M. Richey, "A Cubic Algorithm
for the Directed Eulerian Subgraph Problem," to
appear in European Journal of Operations
Research.
R. Rardin and C.A. Tovey, "Test Travelling
Salesman Problems of Intermediate Complexity."
A.E. Roth and J.H. VandeVate, "Decentralized
Paths to Stability in TwoSided Matching."
D. Solow, R. Stone and C.A. Tovey, "Solving
LCP on Known PMatrices is Probably not NP
Hard."
R. Stone and C.A. Tovey, "The Simplex and
Projective Scaling Algorithms as Iteratively
Reweighted Least Squares Methods."
C.A. Tovey, "Asymmetric Probabilistic
Prospects of Stackelberg Players."
C.A. Tovey, "The Value of Information and
Cooperation in Bimatrix Games: An Average
Case Analysis."
C.A. Tovey, "Simulated Simulated Annealing."
C.A. Tovey, "Simplified Anomalies and
Reduction for Multiprocessor Precedence
Constrained Scheduling."
J.H. VandeVate and J. Wang, "Question
Asking Strategies for Horn Clause Systems."
J.H. VandeVate, "Fractional Matroid Match
ings."
System Optimization Laboratory
Operations Research Department
Stanford, CA 943054022
H. Hu, "On the Feasibilty of a Generalized
Linear Program," SOL 891.
H. Hu, "SemiInfinite Programming," SOL 89
2.
R.W. Cottle, "The Principal Pivoting Method
Revisited," SOL 893.
A.S. Krishna, "Note on Degeneracy," SOL 894.
K. Zikan, "An Efficient Exact Algorithm for the
"LEAST SQUARES" Image Registration
Problem," SOL 895.
A. Marxen, "Primal Barrier Methods for Linear
Programming," SOL 896.
Mathematisches Institut der
Universitat zu K6ln
Weyertal 8690
D5000 K l6n 41
WEST GERMANY
B. Fassbender, "A Sufficient Condition on
Degree Sums of Independent Triples for Hamil
tonian Cycles in 1Tough Graphs," WP 8978.
A. Bachem, A. Dress and W. Wenzel,
"Varieties on a Theme by J. Farkas," WP 8973.
W. Kern, "Verfahren der Kombinatorischen
Optimierung und ihre Gilltigkeitsbereiche," WP
8971.
U. Faigle and W. Kern, "A Note on the
Communication Complexity of Totally Unimodu
lar Matrices," WP 8970.
A. Bachem and M. Niezborala, "Numerische
Erfahrungen bei der Vektorisierunglinearer
Programmierungsalgorithmen," WP 8968.
U. Faigle and W. Kern, "Note on the Conver
gence of Simulated Annealing Algorithms," WP
8967.
U. Faigle, W. Kern and T. Gy6rgy, "On the
Performance of OnLine Algorithms for Partition
Problems," WP 8966.
A. Bachem and A. Reinhold, "On the Com
plexity of the FarkasProperty of Oriented
Matroids," WP 8965.
A. Bachem and W. Kern, "A Guided Tour
through Oriented Matroid Axioms," WP 8964.
M. Hofmeister, "Concrete Graph Covering
Projections," WP 8962.
Mathematical Sciences Technical
Report Series
Department of Mathematical
Sciences
Clemson University
Clemson, SC 296341907
R. Ringeisen and V. Rice, "Cohesion Stability
under Edge Destruction," TR 557.
R. Ringeisen and V. Rice, "When is a Stable
Graph not Stable or Are There Any Stable
Graphs Out There?" TR 557A.
E. Cockayne, B. Hartnell, S.T. Hedetniemi
and R. Laskar, "Efficient Domination in
Graphs," TR 558.
R. Ringeisen and V. Rice, "Cohesion Stable
Edges," TR 559.
R. Ringeisen and C. Lovegrove, "Crossing
Numbers of Permutation Graphs," TR 560.
M. Kostreva, A. Aoun, N. Brown, S. Chatto
padhyay, R. Guidry, T. Ordoyne and R.
Zurovchak, "Linear Complementarity Theory:
1st Generation," TR 561.
B. Piazza and R. Ringeisen, "Connectivity
Generalized Prisms over G," TR 562.
B. Piazza, R. Ringeisen and S. Stueckle,
"Properties of NonMinimum Crossings for
Some Classes of Graphs," TR 564.
B. Piazza, R. Ringeisen and S. Stueckle, "On
the Vulnerability of Cycle Permutation Graphs,"
TR 565.
J. Key and K. Mackenzie, "An Upper Bound
for the pRank of a Translation Plane," TR 566.
D. Shier and N. Chandrasekharan, "Algo
rithms for Computing the Chromatic Polyno
mial," TR 567.
coNTINUES
 
PAGE 5
number twentynine
MARCH 1990
PAGE 6 number twentynine MARCH 1990
S *
Technical Reports &
Working Papers
J. Key and K. Mackenzie, "Ovals in the Design
W(2m)," TR 568.
P. Dearing, P. Hammer and B. Simeone,
"Boolean and Graph Theoretic Formulations of
the Simple Plant Location Problem," TR 569.
M. Kostreva, M. Wiecek and T. Ordoyne,
"Multiple Objective Programming with Polyno
mial Objectives and Constraints," TR 571.
C. Williams, "A KnowledgeBased Approach to
Designing Experiments: Design Expert," TR
572.
V. Rice and R. Ringeisen, "On Cohesion Stable
Graphs," TR 573.
J. Boland, R. Laskar and C. Turner, "On Mod
Sum Graphs," TR 574.
R. Laskar, S. Stueckle and B. Piazza, "On the
EdgeIntegrity of Some Graphs and Their
Complements," TR 575.
R. Laskar, A. Majumdar, G. Domke and G.
Fricke, "A Fractional View of Graph Theory,"
TR 576.
J. Lalani, R. Laskar and S.T. Hedetniemi,
"Graphs and Posets: Some Common Parameters,"
TR 577.
M. Kostreva and M. Wiecek, "Linear Comple
mentarity Problems and Multiple Objective
Programming," TR 578.
G. Isac and M. Kostreva, "The Generalized
Order Complementarity Problem," TR 579.
Research Initiative Program in
Discrete Mathematics and Com
putational Analysis
Clemson University
Clemson, SC 296341907
R. Laskar, R. Rowley, R. Jamison and C.
Turner, "The Edge Achromatic Number of Small
Complete Graphs," URI029.
T. Wimer, "Linear Algorithms on kTerminal
Graphs," URI030.
D.R. Shier, E.J. Valvo and R.E. Jamison,
"Generating the States of a Probabilistic System,"
URI031.
D.R. Shier, "The Monotonicity of Power Means
Using Entropy," URI032.
C. Jeffries, "Fluid Dynamics with Pressure
Diffusion," URI033.
D.R. Shier and G.A. Vignaux, "Adaptive
Methods for Graphing Functions," URI034.
P.J. Slater, "A Summary of Results on Pair
Connected Reliability," URI035.
C.L. Cox, "Implementation of a Divide and
Conquer Cyclic Reduction Algorithm on the FPS
T20 Hypercube," URI037.
R.E. Fennell, "An Application of Eigenspace
Methods to Symmetric Flutter Suppresion,"
URI038.
J.A. Reneke and J.R. Brannan, "Application of
RKH Space Methods to the Filtering Problem for
Linear Hereditary Systems, URI039.
J.A. Reneke and R.E. Fennell, "Canonical
Forms for Distributed Systems Control II," URI
040.
J.A. Reneke and R.E. Fennell, "Convergence of
RKH Space Simulations of Stochastic Linear
Hereditary Systems," URI041.
R.E. Fennell, R.E. Haymond and J.A. Reneke,
"RKH Space Simulation of Stochastic Linear
Hereditary Systems," URI042.
C.R. Johnson and T.A. Summers, "The
Potentially Stable Tree Sign Patterns for
Dimensions Less than Five," URI043.
M.E. Lundquist, "An Implementation of the
Preconditioned Conjugate Gradient Algorithm on
the FPST20 Hypercube," URI044.
S.T. Hedetniemi, R. Laskar, E.J. Cockayne
and B.L. Hartnell, "Efficient Domination in
Graphs," URI045.
S.T. Hedetniemi, M.O. Albertson, R.E.
Jamison and S.C. Locke, "The Subchromatic
Number of a Graph," URI046.
K.R. Driessel, "On Isospectral Surfaces in the
Space of Symmetric Tridiagonal Matrices," URI
047.
P.J. Slater and D.L. Grinstead, "On Minimum
Dominating Sets with Minimum Intersection,"
URI048.
W.H. Ruckle, "Abstract of the Linearizing
Projection, Local Theories," URI049.
W.H. Ruckle, "On WinLose Draw Games,"
URI050.
W.H. Ruckle, "Computer Studies of Coalition
Formation Under Varying Dynamics," URI051.
R.J. Lakin, "State Space Approximation of a
MultimodeComponent System," URI052.
E.O. Hare, S.T. Hedetniemi, R.C. Laskar and
G.A. Cheston, "Simplicial Graphs," URI053.
G.S. Domke, S.T. Hedetniemi and R.C.
Laskar, "Fractional Packings, Coverings, and
Irredundance in Graphs," URI054.
M.M. Kostreva, "Recent Results on Com
plimentarity Models for Engineering and
Economics," URI055.
S.T. Hedetniemi, G.A. Cheston, A. Farley
and A. Proskurowski, "Spanning Trees with
Specified Centers in Biconnected Graphs," URI
056.
J.D. Trout, Jr., "Vectorization of Morphological
Image Processing Algorithms," URI057.
J.P. Jarvis, D.E. Whited and D.R. Shier,
"Discrete Structures and Reliability Computa
tions," URI058.
C.L. Cox, "On Least Squares Approximations to
First Order Elliptic Systems in ThreeDimen
sion," URI059.
G.S. Domke, S.T. Hedeniemi, R. Laskar and
G. Fricke, "Relationships Between Integer and
Fractional Parameters of Graphs," URI060.
W.P. Adams and P.M. Dearing, "On the
Equivalence Between Roof Duality and La
grangian Duality for Unconstrained 01
Quadratic Programming Problems," URI061.
J.A. Reneke and M. Artzrouni, "Stochastic
Differential Equations in Mathematical
Demography: A Review," URI062.
S.T. Hedetniemi and N. Chandrasekharan,
"Fast Parallel Algorithms for Tree Decomposing
and Parsing Partial kTrees," URI063.
B.B. King, "The Dynamics of the Motion of a
Filament: A Survey of the Literature," URI064.
W.H. Ruckle, "A Discrete Game of Infiltration,"
URI065.
R. Geist and S. Hedetniemi, "Disk Scheduling
Analysis via Random Walks on Spiders," URI
066.
~
MARCH 1990
PAGE 6
number twentynine
PAGE 7
Theory of Suboptimal R E V
Decisions
By A. A. Pervozvanskii
and V. G. Gaitsgori
Kluwer, Dordrecht 1988
ISBN 9027724016
This is a research monograph concerned with
large and complex optimization systems, too
large to deal with analytically or numerically. For such
systems the notion of optimality is often dubious, and it
is usually furnished by the systems analyst or the decision
maker rather than the mathematician. In these situations,
insisting by all means on "optimality" may not be justified. An
alternative is to use a more realistic and "relaxed" approach by
exploiting the inner structure of the system, such as the "strong" and
"weak" bonds between its various subsystems. It appears that in
many practical models, after neglecting the weak bonds, the strong
ones recover a simpler system that can be solved by, e.g., aggregation
or decomposition. These solutions are the suboptimall decisions."
Their theory, and the ways of calculating and improving them in
relation to the unknown "optimal" solution of the original complex
problem, is the main theme of this book. A method for improving
suboptimal solutions is the "perturbation method."
Let us illustrate the method in a simple and ideal situation. Consider
a linear program (L, e) : Max (c + cc : (Ao + EA')x = b + eb', x 2 0) for
some e> 0. Suppose that the program is large and complex, so we do
not know its optimal solution x(e) and its optimal value f(e). How
ever, suppose that its "reduced" program (L,0) is easy to solve for x(0).
If this suboptimal x(0) is a unique and nondegenerate basic solution,
and if the solution X(0) of dual of (L,0) is also unique, then for all
"sufficiently small" E > 0 we have the expansion
x(E) = x(0) + EX') + ... + e" x"') + O("^1).
Here the basic components of x00 are given recursively by
(1) = (AB)'(Alx(O) b')
x('1)= (A)1 A1xW, k=1,2,...,m1
while all nonbasic components are zero. (Here A, is a submatrix of A
consisting of the basic vectors.) Also
f(E) = f(O) + E[xT(O)c + )T(0)b' XT(0)A' x(0)] + O(e).
For example: Max{1.3x, x:1.1x, + 0.2x2= 1, x, 0,x2 0} is realization
of Max((1 + 3e)x, x : (1 + e)x, + 2ex2 = 1, x, > 0, x, > 0) at e = 0.1. The
basic part of the exact solution, for small E > 0, is (for m = 2) : XB(e) =
1 e + t2 + 0(E) = 0.90909.... Here the basic part of the suboptimal
solution x,(0.1) = 0.91 is obtained from the above recursive formulae
after solving the simple reduced program: Max(x, x2: x, = 1, x2 0).
I W X S If the primal or the dual are not
unique, then the expansions are
made around particular optimal
solutions obtained after solving
"auxiliary" programs such as Max
xT(O)(c (A1)TX(0)) over all optimal
solutions x(0) of (L,0). Of course,
the above ideas work only if the re
duced program is relatively simple to solve and
if the original program (L,e) is locally stable at
 0 relative to > 0. The perturbation method is formu
lated for linear and then extended to convex programs.
By "stability" the authors mean continuity of the optimal
value function f(e) f(0) and (in the case of uniqueness) x(E)+
x(0) as e* +0. If this fails, then the programs are "singularly
perturbed." For such convex models
Min{f(x) + efl(x) : gO(x) + eg'(x) < 0 = f()
another "auxiliary" program is constructed by adding constraints of
the type vT g(x) < 0, where v is an extreme ray of the unbounded set of
Lagrange multipliers. The new feasible set is now smaller but, since
some "badly behaved" constraints are replaced by "nice" perturba
tions, one may still obtain convergence f(e)f*(0) and x(E) + x*(0) as
e + 0 and reformulate the perturbation method. (The asterisk refers
to the optimal value and the optimal solution of the reduced auxiliary
program.)
The book has six chapters. The first four deal with the s~iboptimal
decisions and the perturbation method for finitedimensional pro
grams. In the last two, the ideas are extended to models that include
differential equations in the constraints. Although the optimal solu
tion is now often available (say in linearquadratic problems of opti
mal control), the perturbation method, in addition to providing infor
mation about the robustness of optimal solutions, allows one to obtain
simpler control designs which are, as a rule, more reliable. The method
also shows how to avoid numerical difficulties if the algebraic Riccati
equation is large. Particularly interesting is a study of the relationship
between singularity and loss of controllability and/or observability.
This reviewer has partly used the book in a graduate course on
optimization attended by mathematicians and operations research
ers. Although the evidence of efficiency of the perturbation method
for nonlinear programs was not convincing after three months of nu
merical experimentation, the students have found the topics provoca
tive and intriguing. (The difficult question of how far one can stretch
e> 0 from e= 0 to retain stability has cropped up repeatedly and,
typically, it could not be answered.) The selection of applications of
both stable and singularly perturbed programs is nonstandard, origi
nal and, in fact, remarkable. It ranges from optimization problems in
input output analysis, interregional transportation problems and
Markov programming (here we find examples of "real life" singularly
cONTINUES
MARCH 1990
Is(OK 
PAE8nme wntieMRH19
perturbed programs) to the engineering examples of suboptimal
regulator syntheses including linear models for continuous techno
logical control problem and ecological system control.
The book is a revision of a Russian text that was published in 1979.
During the last 10 years there has been significant progress made in
parametric optimization that does not appear to have been closely
followed by the authors. No reference has been made in the book to the
school of parametric optimization from and around von Humboldt
University (Nozicka, Bank, Guddat, Klatte, Kummer, Tammer), and
no recent results on sensitivity by several other important contribu
tors (e.g., Fiacco, Gal or Robinson) are mentioned. Had the authors
used pointtoset mappings and lower semicontinuity of the feasible
set mapping in the definition of stability, their presentation would
have been more unified and smoother. Indeed, many of their results
can be readily extended to vector perturbations over "regions of
stability." The English translation is not always precise, and this
occasionally creates ambiguities (e.g., the claim: "For convex pro
grams the existence of a Lagrange vector appears to be sufficient
condition of optimality" or "... there exists an interior point of the
feasible domain, i.e., the Slater conditions are fulfilled..." on p. 38; the
former is a sufficient condition for optimality and the latter claims are
not equivalent).
Prerequisites for reading most of the book are standard undergradu
ate courses in real analysis and linear algebra, plus the essentials of
linear and nonlinear programming. The last two chapters require
some familiarity with control theory. The book is of interest to applied
mathematicians, operations researchers, and electrical engineers, to
both the students and the researchers. In summary, this is an original
and interesting book with many fresh ideas that excite the reader and
reassure him that one can still do useful and mathematically sound,
but not too technical, research in optimization.
The book appears in the new "Soviet Series" Mathematics and Its
Applications program and comes from the IIASA group. This pro
gram is devoted to new emerging (sub) disciplines and their interre
lationships. The idea is to publish books "which are stimulating rather
than definitive, intriguing rather than encyclopedic". The editor
Hazewinkel could not have made a better choice than including this
book in the series.
S. ZLOBEC
Theoretical and Computational Aspects of Simu
lated Annealing
S:y P. J. M. van Laarhoven
I\VI Tract 51, Amsterdam, 1988
during the last several years, simulated annealing (a certain kind of
randomized local search procedure) has become a popular tool for ap
proximately solving large scale discrete optimization problems. The
present book on simulated annealing is, to my knowledge, the second
one that has been written on that topic. The first one, Simulated
Annealing: Theory and Applications, was written by the author of the
present book, together with E. L. Aarts, and published by D. Reidel
Publishing Company in 1987, only one year before the second one.
Thus one may ask whether it was necessary to present a new book on
this subject after such a short period of time. To answer this question,
of course, one has to compare the two.
Firstly, the present tract is a slightly revised version of the author's
doctoral thesis (supervised by J. K. Lenstra and A. Rinnooy Kan),
whereas the first one is a monograph on simulated annealing, written
for a much more general audience (including physisists, electrical
engineers, but not biologists, according to what the authors state in the
preface). As a consequence, the present book by Laarhoven concen
trates on his own results rather than presenting a survey of what has
been done in the field by other researchers. The basic theory of
simulated annealing (conditions for convergence to optimality) is not
treated in depth in either of the two books. Proofs are omitted except
for the simplest version of the convergence theorem in the homogene
ous case. The main concern of the present tract is the (theoretical)
analysis of cooling schedules which do not satisfy the theoretical
conditions for convergence to optimality but are likely to yield good
results in practice. Rules for choosing the cooling schedule are ob
tained by estimating certain parameters of the optimization problem
at hand, using Bayes' Theorem and rather involved mathematical
machinery. Applications and computational results are presented in
order to compare different cooling schedules and to support the
theoretical estimates.
WALTER KERN
Biological Delay Systems: Linear Stability Theory
Cambridge Studies in Biology: 8
by N. MacDonald
Cambridge University Press, Cambridge, 1989
ISBN 0521340845
This excellent book gives a survey on the most important methods and
results of the theory of delay equations arising from biological control
systems. The special structure of the considered delayed differential
equations is very well motivated by striking examples. A large part of
the book is dedicated to mathematical modelling and description of
the following biological systems: neurophysiology of the retina, insect
maturation times, maturation of blood cells, population models, incu
bation times in the epidemiology, Neuron interaction, chemostat
models like the Monod model and others. These models are not an a
posteriori justification for the development of an abstract "oversized"
theory but the apriori motivation for the development of mathemati
cal tools that work in the concrete situations arising from mathemati
  ~~s~~
PAGE 8
number tiventynine
MARCH 1990
0A
BGDK
R K V I E W S
cal biology. The concept of the book is influenced bythe The answer is well known; instead of proving a pro
work of K. L. Cooke. The intended application in bio gram once it has been written, start its design from its
logical control does not require the embedding of delay proof. The construction of a loop begins with the proposal of
equations in the wider class of general functional differential a loop invariant. Several classical examples are proposed to il
equations that one finds in the work of J. K. Hale. lustrate this method. The resulting programs are analyzed and, in
The book is essentially concerned with local, or linear, stability analy some cases, improved along the line suggested by the analysis. The
sis. This requires the linearization about a fixed point and the study of authors give a good bibliography of previous works on this program
the roots of a certain algebraic equation, the socalled characteristic ming method.
equation of the system, which is analogous to the characteristic equa The effect of assignments are discussed, introducing the concepts of
tion in ordinary differential equations. The stability analysis requires weakest preconditions or strongest postconditions. The effects of
the localization of roots with negative real parts. This motivates the indexing and pointers are discussed. Procedure calls are taken into
intensive study of Hurwitz polynomials. In the field of discrete delay, account. This is a very deep review of essential programming features.
where the change of the state of the system at time t is affected by In the second part of the book recursion is considered. Induction is the
terms depending on tT, the book treats first and secondorder keyword. The authors show how a recursive procedure can be proven
systems as well as higherorder systems and systems with two delays. to be correct by induction and how it can be analyzed, the recursive
Moreover, one finds an extended section on distributed delay where procedure giving an inductive or implicit definition of the number of
the present state of the system at time t is affected by an integral over operations which can be solved using ordinary mathematical ways.
all past states of the system. One also should mention the independent We have here the same situation as with iterative programming. It is
and commensurate delays as well as reducible delays and the role of better to start with the proof of the program (an inductive relation on
linear subsystems. the function to compute, a special case of explicit definition) to build
The motivation and the presentation of the mathematical topics are of the program rather than first writing it, then trying to prove it.
excellent clarity. Every chapter is followed by many informative exer Pertinent nontrivial examples are given to illustrate this method.
cises, with solutions given in an appendix. Thus, this volume will be
an excellent textbook for any graduate course on delay equations, and
it easily can be used for private studies. In general, this outstanding
book can be recommended strongly for students at the graduate level
and for research workers in mathematical biology and control theory.
J. WEYER
AlgorithmsThe Construction, Proof, and Analy
sis of Programs
by P. Berlioux and P. Bizard
Wiley, Chichester, 1987
The concern of this book is to help the reader to write correct pro
grams. In every science, it is well known that any number of examples
will never prove a theory; testing a program may show that it contains
bugs, never that it is correct. There is a unique way to guarantee the
value of a program, to prove that it is correct from its text (it will
remain to guarantee that no typing error has been made when
entering it in the computer).
The first chapter gives the theoretical foundations of formal program
proofs, using Floyd's inductive assertions and Hoare's formalization.
The authors emphasize the fact that the result is not really satisfactory,
a lot of derivations for a very short and trivial program! The proof
could be simplified using less formal tools, as it is usually done in
mathematics. But even with less convincing proofs, the problem
remains to discover a loop invariant to start the proof.
Finally, the authors consider one of the methods which have been
proposed to transform recursion into iteration and its use to derive
iterative programs from recursive definitions, with some efficiency
concerns.
This book has been written for people having some experience in
programming. It emphasizes the fact that program construction starts
with the proposal of what can be called "a recurrence hypothesis"
(loop invariant for iterative programming, recurrence relation for
recursive programming). It does not consider where this hypothesis
comes from. Like most programming books, building a program
starts with something like, "Assume that we have been able to do
that...." The questions remain: Why such a choice? Are you sure it is
the best? How did you find it? This is especially clear with towers of
Hanoi, "If n> 1, we begin by transferring a tower of height n1 from rod
A to rod C, using B as an intermediary." Why do you do that? Without
an answer to this question, it is not clear that the resulting program is
the best possible one.
Nevertheless, it is clear that the most difficult disk to move is the
biggest one. It can be moved only if there is no other disk on it, so the
n1 other disks must be somewhere else, on the two other rods. If so,
we can take this disk off of its rod, but we can put it on another rod only
if this rod is empty. Thus the n1 other disks must be on the third rod,
and there is no other possible solution. A slightly different presenta
tion gives a quite different result. The same remark is true for some
other programs of the book. For instance, the binary search can be
written in a simpler way, without extra tests before entering the loop.
coNNUnE
~3aa~s~
    ~ 
MARCH 1990
number twentynine
PAGE 9
Pen
This book is not a book on how to invent a new program. It is a good
textbook on some methods now available to construct correct and
efficient programs, assuming that you have some idea of the method
to be implemented. It is clear, wellwritten, welldocumented and
deep enough. There are many examples which are not all toy ex
amples. This book is certainly worthwhile for any programmer or
informatics student who wishes to know what is meant by program
correctness, program efficiency or transformation of recursion into
iteration without going into a lot of unnecessary details. It will not
provide all the ways presently available to build, analyze, transform,
or improve a program, but it will introduce the reader to this wide new
area, the science of programming.
JACQUES ARSAC
Introduction to Optimization
by E. M. L. Beale
Edited by L. Mackley
Wiley, Chichester, 1988
ISBN 0471917605
"The book is based upon a series of lecture notes written by Professor
E. M. L. Beale for his undergraduate course Introduction to Optimi
zation,' given at Imperial College where he was a Visiting Professor in
Mathematics from 1967 until his death in December 1985." /Lynne
Mackley./
In this book the emphasis is put more on methods, algorithms and the
practical problems of why and how methods succeed or fail, rather
than on deep theory and rigorous proofs. As it contains different
applications of optimization in industry and many suggestions for
choosing efficient numerical procedures for solving concrete optimi
zation problems, this book could be highly recommended to engi
neers having to solve practical optimization problems. Moreover, it
would be a very good introduction for beginners in this area.
The book is divided into three parts. The first part, "Unconstrained
Optimization" (chapters 24), is concerned with the main techniques
for solving problems both for functions of onevariable and multi
variable. Chapter 2 starts with considering the unconstrained optimi
zation of functions of continuous variables and reviews iterative and
valleydescending methods.
Chapter 3, Onedimensional Optimization, describes a popular ap
proach to solving this problem, i.e. Newton's method, the bisection,
the method of false position and its modification. A Wijngaarden's
method which uses both linear interpolation and bisection is
recommended. Finally, some insight into the sort of ideas that go into
development of optimization algorithms is given. Chapter 4 is about
multidimensional optimization. It clearly explains why there is no
single method which is effective for all problems of this type, resulting
in a vast literature on numerical methods for unconstrained optimiza
tion of functions of n variables. Most of this chapter applies to finding
local optima of functions that are twice differentiable.
The methods used to solve linear programming problems and appli
cations of linear programming in industry are treated in the second
part "Constrained Optimization: Linear Programming" (chapters 5
9). Chapters 5 and 6 describe the simplex method for linear program
ming and give the basic steps of the revised simplex method (also
known as the inverse matrix method). The efficient modelling and
systematic documentation of linear programming problems are re
viewed here. Duality and its applications, dual simplex method and
examples are studied in detail in Chapter 7. This chapter also deals
with parametric programming in two basic forms, variation of the
objective function and variation of the righthand side. Chapters 8and
9 answer the question of how to apply linear programming problems
in industry.
Nonlinear, discrete and dynamic programming are the topics of the
third part (chapters 1012). Chapter 10 is concerned with nonlinear
programming. It starts by reviewing the basic theory of optimization,
i.e. Lagrange multipliers and the KuhnTucker conditions and then
presents two methods which are in practical use. The first is based on
the concept of separable programming and the other on the reduced
gradient method. Classical approaches to integer programming and
its potential difficulties are discussed in Chapter 11. The cutting plane
method and the branch and bound methods are discussed as general
strategies, and a simple example makes the presentations clear.
Chapter 12 discusses dynamic programming as an alternative to
integer programming for some problems in combinatorial
optimization. The art of dynamic programming formulation is illus
trated with a shortestroute problem. In general, this part shows how
the techniques of linear programming can be applied to solve nonlin
ear and discrete optimization problems.
ANNA RYCERZ
_I_ I ~I I 1_1 ~ __ ~
MARCH 1990
PAGE 10
number twentynine
number twentynine
MARCH 1990
SRaeLR
Volume 44, No.3
E.J. Anderson and A.S. Lewis, "An Extension of
the Simplex Algorithm for SemiInfinite Linear
Programming."
J.M.Y. Leung and T.L. Magnanti, "Valid Ine
qualities and Facets of the Capacitated Plant
Location Problem."
R. Chandrasekaran and A. Tamir, "Open
Questions Concerning Weiszfeld's Algorithm for
the FermatWeber Location."
I. Adler, N. Karmarkar, M.G.C. Resende and G.
Veiga, "An Implementation of Karmarkar's
Algorithm for Linear Programming."
E. Balas, J.M. Tama and J. Tind, "Sequential
Convexification in Reverse Convex and Disjunc
tive Programming."
M.W. Jeter and W.C. Pye, "An Example of a
Nonregular Semimonotone QMatrix."
DZ. Du and XS. Zhang, "Global Convergence of
Rosen's Gradient Projection Method."
Volume 46, No.1
R.H. Byrd, C.L. Dert, A.H.G. Rinnooy Kan and
R.B. Schnabel, "Concurrent Stochastic Methods
for Global Optimization."
E.S. Gottlieb and M.R. Rao, "The Generalized
Assignment Problem: Valid Inequalities and
Facets."
E.S. Gottlieb and M.R. Rao, "(l,k)Configuration
Facets for the Generalized Assignment Problem."
Y. Ye, "A 'BuildDown' Scheme for Linear Pro
gramming."
B. Kalantari, "Karmarkar's Algorithm with
Improved Steps."
C. Roos, "An Exponential Example for Terlaky's
Pivoting Rule for the CrissCross Simplex
Method."
H. Hu, "A OnePhase Algorithm for SemiInfinite
Linear Programming."
K.C. Kiwiel, "Proximity Control in Bundle
Methods for Convex Nondifferentiable Minimiza
tion."
Volume 46, No.2
P. Tseng and D.P. Bertsekas, "Relaxation
Methods for Monotropic Programs."
C.L. Monma, B.S. Munson and W.R.
Pulleyblank, "MinimumWeight TwoConnected
Spanning Networks."
M.D. Asic, V.V. KovacevicVujcic and M.D.
RadosavljevicNikolic, "Asymptotic Behaviour of
Karmarkar's Method for Linear Programming."
R. Ge, "A Filled Function Method for Finding a
Global Minimizer of a Function of Several
Variables."
M.I. Henig, "Value Functions, Domination Cones
and Proper Efficiency in Multicriteria Optimiza
tion."
R. Chandrasekaran and A. Tamir, "Algebraic
Optimization: The FermatWeber Location
Problem."
F. Granot and J. Skorinkapov, "Towards a
Strongly Polynomial Algorithm for Strictly
Convex Quadratic Programs:An Extension of
Tardos'Algorithm."
M. Meanti, A.H.G. Rinnoy Kan, L. Stougie and C.
Vercellis, "A Probabilistic Analysis of the Multi
knapsack Value Function."
R. Hettich and G. Gramlich, "A Note on an
Implementation of a Method for Quadratic Semi
Infinite Programming."
J. Rohn, "A Short Proof of Finiteness of Murty's
Principal Pivoting Algorithm."
Application for Membership
Mail to:
THE MATHEMATICAL PROGRAMMING SOCIETY, INC.
c/o International Statistical Institute
428 Prinses Beatrixlaan
2270 AZ Voorburg
The Netherlands
Cheques or money orders should be made payable to
THE MATHEMATICAL PROGRAMMING SOCIETY, INC.
in one of the currencies listed below.
Dues for 1990, covering subscription to volumes 4648 of
Mathematical Programming, are Dfl.115.00 (or $55.00 or
DM100.00 or 32.50 or FF345.00 or Sw.Fr.86.00).
Student applications: Dues are onehalf the above 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 subscription is for
my personal use and not for the benefit of any library or institution.
I enclose payment as follows:
Dues for 1990
Name (please print)
Mailing Address (please print)
Signature
Faculty verifying status
Institution
PAGE 11
I ~
PG12numberwn M
C ONTRIBUTIONS for a Mixed Integer Programming library
of test examples are invited by Peter Conrad and Prof. R.C.
Daniel, Univ. of Buckingham, Buckingham, MK18 1EG, Eng
land, Telephone: (0280) 81408. !Prof. Dr. Reiner Horst, Univer
sitit Trier, Postfach 3825, 5500 Trier, FRG, has issued a call for
papers for a new Journal of Global Optimization to be published
beginning in 1991. JAIRO'90, the annual Conference of the Op
erational Research Society of Italy, will be held in Sorrento, Oct.
35, 1990. Contact Prof. Antonio Sforza or the Secretariat, Insti
tuto di Fisica, Matematica e Informatica, Facolta di Ingegneria,
University di Salerno84084 Fisciano (Salerno), Tel. +3989
822233/822424. !JDeadline for the next OPTIMA is June 1,1990.
Books for review should be
sent to the Book Review Editor,
Prof. Dr. Achim Bachem,
Mathematiches Institute der
Universitiit zu Kiln,
Weyertal 8690, D5000 Kiln,
West Germany.
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.
Elsa Drake, DESIGNER
P T I M A
MATHEMATICAL PROGRAMMING SOCIETY
303 Weil Hall
College of Engineering
University of Florida
Gainesville, Florida 32611 USA
FIRST CLASS MAIL
~11~~18181~'sC"""""~IBIICI~""""
II. I 1. II~~~ ~~~~~~
MARCH 1990
PAGE 12
number twentynine
