| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The tutorial will illustrate, in manufacturing and service settings, how OR can enhance the quality of operations. It will adopt a broad, international perspective, and demonstrate that OR can be successfully integrated into corporate strategy to achieve world class performance. It will illustrate relationships between OR and other company functions.
The tutorial will also demonstrate how OR, with its multi-disciplinary approach and practical approach can be aid in the quest for management excellence. Through the analysis of case studies, attendees will be able to exercise their own judgement as to the value of the World Class Operations philosophy.
Peter Stuckey
Department of Computer Science
University of Melbourne
Parkville VIC 3052
Australia
Constraint programming languages are a recent development in computer science, originating in research conducted in the late 1980's at Melbourne and Monash University, ECRC in Munich and the University of Marseille. The first constraint programming languages, called constraint logic programming (CLP) languages, extended logic programming languages by replacing unification by a test for constraint satisfaction using a dedicated constraint solver.
Since that time constraint programming has been used in many diverse areas and in particular for tackling difficult multi-faceted combinatorial problems. Such problems are difficult to express and solve in conventional programming languages. Typically, either an off-the-shelf solver is used or a dedicated solver is developed in a language such as FORTRAN or C. Modern constraint programming languages combine these approaches by giving a high-level modelling language which is a true programming language built on top of a constraint solver.
In the ten years since CLP was invented, constraint programming languages have matured, and commercial languages and toolkits are now appearing in the market place amid substantial hype. In this tutorial we will cut through this hype and identify the key features of modern CLP languages and explain their advantages and limitations. In particular, we will:
Composite concave programming (c-programming for short) is a relatively new nonlinear parametric optimization method designed for solving nonlinear combinatorial minimization problems whose objective functions have a composite concave structure. The parametric nature of c-programming enables it to extend the scope of operation of basically any optimization method in a systematic and uniform manner.
In this tutorial we shall examine the mathematical idiom of c-programming, the general nature of the algorithms that it forges, some of its natural and not-so-natural applications, as well as technical issues concerning the incorporation of c-programming topics into the Operations Research/Management Science/Mathematical Programming undergraduate curriculum. It will be demonstrated that although c-programming is based on fundamental concepts and tools of classical optimization theory, it is particularly suited for nonlinear combinatorial optimization problems as well as a large class of global optimization problems.
It will be demonstrated that c-programming provides a uniform and systematic framework for the analysis and solution of a number of well know generic problems such as fractional programming problems, convex multiplicative problems, variance minimization problems, mean-variance ratio problems, and a large class of dc programming problems.
John A. George Tabu search offers useful strategies for solving combinatorial and nonlinear optimization problems by utilizing an adaptive memory design that is more flexible than the memory designs available to branch & bound (B&B) and branch & cut (B&C) procedures. Widespread applications of such strategies have been developed for special problem classes, often resulting in "new records" for solving problems in areas ranging from industrial planning to environmental control to biomolecular design. We show this adaptive memory also offers useful decision strategies for solving general (as well as "special class") problems from the domain of pure and mixed integer programming (IP). Procedures based on tabu branching, scatter search and star-paths, are shown to be capable of probing the solution space in ways that are inaccessible to B&B and B&C methods. Accompanying this, we identify a set of theorems that enhance the effectiveness of this probing. These results have a highly visual component that allows their content to be communicated effectively by diagrams. The power of the approaches compared to previous approaches can also be shown graphically. These results additionally make it possible to integrate tabu search with branching and cutting methods, thereby yielding new procedures -- both heuristic and exact -- for solving general IP problems. Nicola Ward Petty Rikke Ravnborg OBJECTIVES:The aim of the tutorial is to introduce the basic concepts and principles of multimedia teaching, discuss its advantages and pitfalls, and let the participants experience an implementation of this technology on an interactive basis. TOPICS COVERED: Introduction to multimedia computer systems; Principles of multimedia instruction; Contrast to conventional lecture-style teaching; Advantages, disadvantages, and pitfalls of multimedia instruction Brief introduction to MENTOR system; Demonstration of principles of multimedia instruction using MENTOR Experience with developing the 'systems thinking and methodology module' for MENTOR; Hands-on experience with MENTOR for participants. James Powell This tutorial will provide an up-to-date presentation of the Theory of Constraints methodology, which has developed considerably in its brief history, and continues to develop, with the Critical Chain method being the most recent development. The tutorial will start with a brief history of the TOC, and trace its development from a manufacturing method through to the generic problem solving methodology, the Thinking Processes. It will put into context the specific applications, to manufacturing, distribution and marketing, and project management, that have been featured in Eli Goldratt's novels, The Goal, It's Not Luck, and Critical Chain, respectively.The presenter will describe the Thinking Processes, and give examples to show how they can be used together and separately. These tools effectively provide the means to resolve conflicts, give criticism in a creative and supportive way, decide on an appropriate strategy for an organisation, and determine a robust way of achieving strategic goals.The tutorial will also provide an introduction to the new Critical Chain methodology, which builds on the well known Critical Path methodology, taking into account resource constraints that often cause projects to run late and over budget. Discussion will focus on the use of critical chain project scheduling and project management on projects in Australia and New Zealand. Finally, the use of project management software tools to support the critical chain methodology will be discussed. Andrew Gale In recent times the private health insurance system in Australia has been the subject of review and modification. The effect of the four categories of membership, introduced with effect from 1 October 1996, is not well understood by the public. A model has been constructed to evaluate the financial effect of actual and proposed changes to the system on various sections of the community. The key variables used by the model are age, sex, and category of membership. The role and importance of warranty has changed significantly over the last two decades. Currently all most all products (consumer durables, commercial and industrial) are sold with some form of warranty attached. Warranty is of importance to both manufacturers and consumers. From the consumers point of view, warranty provides some information about product reliability and quality and acts as an insurance in the event of early failure of the product. From the manufacturer's view point, it protects the manufacturer from unreasonable claims from the consumers and can be used as a signal to market the product. There are many different aspects to warranty and these have been studied by researchers from diverse disciplines such as engineering, accounting, marketing, statistics, economics, law and operations research. A variety of stochastic models have been developed to study the different aspects. Selling a product with warranty results in additional costs to the manufacturer due to the servicing of warranty. These costs vary from < 1% to > 5% of the sale price depending on the industry. As such, the total warranty costs to manufacturers run into billions of dollars. Operations research can help reduce this cost through realistic modelling of the warranty process and proper model analysis. The tutorial will cover the following topics: The presentation will be at an introductory level and aimed at researchers and practitioners in operations research. It will deal with both theoretical and application issues and will be highlighted through some real case studies. Matteo Fischetti Crew Management is concerned with building the work schedules (rosters) of crews needed to cover a given set of timetabled trips.We give a general definition of the problem, and showpossible operational constraints and objective functions arisingin real-world applications (concerning mass transit companies, airline companies, and railway companies). Several mathematical programming and graph-theory formulations, and different exact and heuristic algorithms proposed in the literature are presented. Advantages and disadvantages of the proposed formulations and algorithms are discussed. In practice, the overall problem is decomposed into two subproblems, called Crew Scheduling (CSP) and Crew Rostering (CRP).In CSP the short-term schedule of the crews is considered, and a convenientset of duties covering all the trips is constructed. In CRP the selected duties are sequenced to obtain the final rosters.Two main solution approaches are illustrated in more details.CSP is solved by generating a very large number of potential duties, and by solving a Set Covering Problem} to select a minimum-cost set of duties covering all the trips. CRP is solved by using a Lagrangian relaxation of a graph-theoretical formulation of the problem. Extensive computational results on real-world instances arising in theItalian railway company, Ferrovie dello Stato SpA, are presented. We survey heuristic search techniques for the solution of combinatorial optimization problems. We first introduce basic constructive approaches, such as greedy and greedy randomized algorithms, which are used to build feasible solutions. Next, we show how solutions generated by such approaches may be successively improved by local changes. We present typical neighborhood definitions for some classes of combinatorial optimization problems and basic local search procedures such as iterative improvement, steepestdescent, and other recent, more advanced strategies such as variable neighborhood search and large step optimization. We discuss the limitations of both constructive and local search approaches. We show how they can be combined with the use of principles and ideas such as efficient data structures, preprocessing, reinitialization, controlled randomization, search intensification, and search diversification, leading to new classes of heuristics which considerably improved the efficiency of approximate algorithms and enlarged the size of solvable instances of difficult problems found in practice. These new approaches are generally named as metaheuristics, in the sense that the process of finding good solutions most often consists in applying at each step a subordinate heuristic, which has to be designed and tailored for each particular problem. We review several classes of metaheuristics, such as simulated annealing, tabu search, greedy randomized adaptive search procedures, genetic algorithms, and asynchronous teams. We describe their main characteristics, and we comment on common issues, specific features, advantages, and drawbacks. We also discuss in details parallelization issues and strategies for each of these approaches.
Department of Management
University of Canterbury
Private bag 4800
Christchurch New Zealand
Presenters:
Tutorial 5 - TD1 Multimedia Teaching of OR
Department of Management
University of Canterbury
Private Bag 4800
Christchurch
New Zealand
University of Strathclyde
Glasgow
UK
GENERAL COMMENTS:Multimedia teaching will without doubt become more widespread as a vehicle for university-level instruction. It is particularly suitable for technical and quantitative topics where visualization of graphical representations enhances understanding of what is going on. OR, with statistics and other quantitative subjects, lends itself well to this mode of teaching. It is therefore important for instructors in OR to become familiar with multimedia teaching. The tutorial will make use of the Learn-OR package of the MENTOR system to demonstrate the principles of multimedia instruction and its implementation.
Presenter:
Tutorial 6 - TA1: Overview of the Theory of Constraints
Avraham Y Goldratt Institute, Australasia
CPO Box 1639
Auckland
New Zealand
Presenters:
Tutorial 7 - TB1: Modelling Health Insurance Outcomes in Australia
Alan Brown
Melbourne
Australia
National Mutual Health Insurance
Melbourne
Australia
Presenters:
Tutorial 8 - TC1 : Modelling and Analysis of Warranties
D.N. Pra Murthy
Department of Mechanical Engineering
The University of Queensland
St Lucia, Q 4072
Audtralia
murthy@mech.uq.edu.au
Presenters:
Tutorial 9 - WB7.1: Models and Algorithms for Crew Management
Alberto Caprara, Paolo Toth, and Daniele Vigo
DEIS, University of Bologna
Bologna
Italy
paolo@promet4.deis.unibo.it
DMI, University of Udine
Udine
Italy
Presenter:
Tutorial 10 - WB7.2: Local Search and Methaheuristics
Celso C. Ribeiro
Catholic University of Rio de Janeiro
Department of Computer Science
R. Marques de Sao Vicente 225
Rio de Janeiro 22453-900
Brazil
celso@inf.puc-rio.br