21331201

Embed Size (px)

Citation preview

  • 7/28/2019 21331201

    1/3

    Lev: Book Reviews274 Interfaces 36(3), pp. 272278, 2006 INFORMS

    interesting and unusual and are likely to appeal to

    applied students. They are designed to test the con-

    cepts. The authors provide answers to many of the

    problems.To sum up, the authors give a good account of

    optimization techniques. The irritating typographical

    errors are a definite drawback. I think the authors

    should correct the errors and bring out a revised edi-

    tion. At $79.95, the book would be quite expensive

    for Indian students, and a less-expensive paperback

    edition would be welcome.

    References

    Sharma, J. K. 2003. Operations Research: Theory and Applications.

    Macmillan India Ltd., New Delhi, India.

    Taha, Hamdy A. 2002. Operations Research: An Introduction. Prentice-Hall, Inc., Upper Saddle River, NJ.

    M. Z. Anis

    Indian Statistical Institute, 203 B. T. Road,

    Calcutta 700108, India, [email protected]

    Maros, Istvan. 2003. Computational Techniques ofthe Simplex Method. Kluwer Academic Publishers,

    Dordrecht, The Netherlands. 325 pp. $139.00.

    In Computational Techniques of the Simplex Method,

    Istvan Maros, a Hungarian mathematician at the

    Imperial College of London, describes the simplex

    method and its implementation in real problems.Andrs Prkopa, one of his professors, provides an

    introduction highlighting the books features and

    virtues. Maros has an original outlook on the simplex

    method, which helps him explain the difficulties of

    applying it in real cases and the computational tech-

    niques for specific linear programs.

    The book is aimed at professors and students who

    want to look deeply into the computational tech-

    niques of the simplex method in linear programming

    (LP). Researchers can also benefit, finding new proce-

    dures for constructing new algorithms in other areas

    of operations research. Hence, I highly recommend

    Maross book as a reference for graduate students in

    computing and linear programming.

    In the first part of the book, Maros describes the pre-

    liminaries and main aspects of linear programming;

    in the second part, he focuses on the computational

    aspects of the simplex method. Three chapters con-

    stitute the first part: the main elements of linear pro-

    gramming (Chapter 1), the essentials of the simplexmethod (Chapter 2), and large linear-programming

    problems (Chapter 3). The second part contains eight

    chapters on computational issues in implementingthe simplex method, including the characteristics oflinear-programming software (Chapter 4), a techni-cal description of data structures and basic opera-

    tions (Chapter 5), the standard linear-programmingproblem and its mathematical programming systems

    (MPS) format (Chapter 6), and the preprocessing pro-cedure for linear programs before the simplex method(Chapter 7). In Chapter 7, the author explains ways

    of reducing the size of the problem, improving itsnumerical characteristics, and detecting infeasibility

    or unboundedness without solving the problem. He

    also explains sparsity reduction of a matrix of coef-ficients and problem scaling, which are essential to

    solving large linear programs. In Chapter 8, Maroscovers algebraic techniques for LP matrices, such as

    basis inverse and matrices factorization, triangularmatrices, and lower and upper triangular factoriza-tion.

    Chapters 9 and 10 are longer and more complexthan the preceding chapters. In Chapter 9, Maros

    devotes 100 pages to explaining the primal method,highlighting computational procedures for solving

    linear programs. He discusses different ratio tests,

    handling degeneracy, numerical stability, memoryrequirements, and computational effort. He quotes

    from Dantzig (1963) extensively. He also describesWolfes algorithm and the Devex pricing method and

    real cases of problems related to tolerance, pertur-bation, and shifting. In Chapter 10, he covers thedual algorithm, including such topics as dual feasibil-

    ity, improving dual solutions, extended general dualalgorithms, degeneracy, and the dual Devex pricing

    method.Finally, Chapter 11 is devoted to miscellaneous

    aspects of implementing the simplex method. Maros

    also summarizes some of the main points of the previ-ous chapters and suggests further topics for research,

    such as tuning parameters, performance evaluation,and alternatives to the simplex method.

    The book ends with a list of references and a short

    glossary. A more extensive glossary might have beenuseful.

    This interesting book about linear programmingand related computational techniques should be use-

  • 7/28/2019 21331201

    2/3

    Lev: Book ReviewsInterfaces 36(3), pp. 272278, 2006 INFORMS 275

    ful to graduate students and researchers who wish to

    learn the key concepts of the simplex method. Many

    readers, students, and researchers can update their

    knowledge about linear programming by studyingMaross book.

    Reference

    Dantzig, G. B. 1963. Linear Programming and Extensions. Princeton

    University Press, Princeton, NJ.

    Javier Faulin

    Department of Statistics and Operations Research, Public

    University of Navarra, 31006 Pamplona, Spain

    Tkindt, Vincent, Jean-Charles Billaut.2002. Multicriteria Scheduling: Theory, Models and Algo-

    rithms. Springer-Verlag, Berlin, Germany. 303 pp.$107.00.

    Scheduling problems are hard. They are even

    harder when multiple objectives must be considered.

    Improving customer service, maximizing resource uti-

    lization, and reducing inventory are among the many

    important objectives that one faces when trying to

    solve a scheduling problem. Choosing a way to

    address the multiple objectives is the first difficult

    step. You could form a new objective that is a linear

    combination of the objectives or optimize one objec-

    tive while satisfying bounds on the others. You could

    minimize the distance from an ideal (but unreach-able) solution or optimize the objectives in sequence.

    You could resort to finding any solution that satis-

    fies constraints on the objectives or find the complete

    set of solutions that are not dominated by any other

    solution.

    The literature on scheduling problems with a single

    objective is extensive, whereas published academic

    research on scheduling problems with two or more

    objectives is not. However, it is an ever-growing body

    of knowledge. In Multicriteria Scheduling, Tkindt and

    Billaut have carefully organized this material for those

    who are studying this complex topic and for those

    who need to locate an algorithm or review the com-

    plexity of a particular problem.

    The book, translated from French by Henry Scott,

    begins with introductory material on scheduling and

    multicriteria optimization. The scheduling material is

    adequate and similar to that found in other texts,

    including Brucker (2004) and Pinedo (2005). In the

    chapter on multicriteria optimization, the authors,

    who are faculty members at the University of Tours,

    provide a systematic overview that covers a wide

    range of relevant topics. They precisely define dif-ferent types of Pareto optimal solutions and discuss

    different techniques, including lexicographical orders,

    interactive methods, and goal programming.

    The chapter on multicriteria scheduling problems

    includes a useful classification of solution approaches.

    When faced with such problems, decision makers

    may want a unique solution that balances the various

    criteria. In this case, they must provide parameters

    that can be used to establish the relative priority of the

    objectives. Alternatively, decision makers can partici-

    pate in the search process, in which case they look at

    each solution generated and select new search direc-tions until they obtain a satisfactory solution. Finally,

    decision makers may want to obtain a set of nondom-

    inated solutions from which to choose a solution.

    The authors have organized the book into chapters

    covering different types of multicriteria scheduling

    problems: single-machine problems, parallel-machine

    problems, flow-shop problems, job-shop problems,

    open-shop problems, and hybrid flow-shop prob-

    lems. They formulate particular scheduling problems,

    present exact algorithms and heuristics, give exam-

    ples to help explain the algorithms, and discuss com-

    putational complexity results. The problems cover a

    wide variety of combinations of objectives, includ-

    ing some for which the authors have previously pub-

    lished original results. The authors provide citations

    for all of the results, enabling interested readers to

    investigate selected topics further.

    In the appendices, the authors review the notation

    and present summary tables that classify multicrite-

    ria scheduling problems according to their complex-

    ity. The 18 pages of references span over 50 years of

    academic research into scheduling problems.

    Multicriteria Scheduling is a unique resource. It is

    similar in spirit to Peter Bruckers 2004 text, Schedul-

    ing Algorithms, in that both books are concerned

    with organizing scheduling problems using the tra-

    ditional three-field classification scheme, describing

    their complexity, and presenting solution algorithms.

    This book is a useful complement to Scheduling Algo-

    rithms because Brucker does not cover problems with

    multiple criteria extensively. Likewise, Leungs (2004)

  • 7/28/2019 21331201

    3/3