For more information, please contact:

Dr. Lutfar Khan (tel: 9688-4687 or khan@matilda.vut.edu.au) orProf. Bob Johnston (tel: 9905-3422 or rejohnston@eng.monash.edu.au)

Expert Systems in Banking

For details contact: Simon Goss 9626 7274, Ross Gayler 9643 8853, Julie Gledhill 9818 0795.

Use of Distributed Interactive Simulation for Naval Training

Abstract

The Royal Australian Navy (RAN) plans to use Distributed Interactive Simulation (DIS) to enhance its team training capability. Initially it is planned to link trainers at the same location and later connect these to other external trainers and live assets both in port and at sea. Linking manned simulators helps to more closely imitate operational environments with different platforms (ships/aircraft) interacting and will provide more sophisticated and effective command team training for the RAN s surface warfare fleet. DIS will also allow these disparate simulators to be linked to the outside world for a wider range of tactical training. Integration of the trainers will also provide the ancillary benefits of course length reductions, system manning efficiencies, and increasing the training environment fidelity.

This paper discusses the application of DIS technology to the RAN s surface warfare operations room simulators. A general overview of DIS and other US modelling and simulation initiatives will also be provided.

Regulating agencies need to make choices on where to inspect trucks carrying hazardous goods. If the capacity of a facility to inspect is very large then locating the facilities to inspect as many trucks as possible becomes a m-cover problem. In the more practical case where the capacity is limited, we formulate a new location problem. We study its complexity and develop some heuristics and bounds for this problem. We will also introduce the concept of "inspection equilibrium" and providesome preliminary results on computing it.

Short Bio:

Pitu Mirchandani is a Professor and Department Head of Systems and Industrial Engineering at the University of Arizona. He has BS and MS degrees in Engineering from UCLA, and adoctoral degree in Operations Research from MIT. His researchinterests are in logistics, location, scheduling, and systems designwith applications in production, transportation, and manufacturing.He has published over 50 papers and two books on locationtheory.

If people would like to meet with Pitu separately, could they please ring me or email me their requests...

Thanks.

--Mohan

Lecture Series in

CONVEX ANALYSIS

Any enquiries should come to me, phone 9344 6761, or e'mail craven@mundoe.maths.mu.oz.au

Dr. B. D. Craven
1 May 1996

Parking is not available on campus without a permit. Perhaps yourinstitution can give you a University of Melbourne parking permit.

Business Modelling in Excel

A Stochastic Modelling Problem in a High-Tech Steel Production Plant

by
Dr A H Christer
Centre for Operational Research and Applied Statistics
University of Salford

2-3pm Wed, 15 May 1996
RACEVIEW , VUT Footscray

The presentation reports on a maintenance modelling actionresearch study of the roll change equipment in a high volume steelmill. The equipment is of a "preparedness" variety, in that defectsare only recognisable when the equipment is required for use. Astochastic model of behaviour of the plant under various servicemaintenance systems is developed, and the consequences to productiontime of alternative service periods and service quality levelspredicted. The increased value in terms of insight given tomanagement of stochastic modelling over simpler options ishighlighted.

For further info - contact Peter Cerone - pc@matilda.vut.edu.au

Modelling and Quality of Automatic Quality Checks

by
Dr A H Christer
Centre for Operational Research and Applied Statistics
University of Salford

2-3pm Thur, 16 May 1996
RACEVIEW , VUT Footscray

High speed high volume automatic production techniques arecapable of producing defective as well as quality products. Toreduce the risk of the former, monitoring checks are commonlyintegrated wit hin a production process to both identify defectiveproduction and to monitor overall quality performance.

The paper addresses the problem of assessing the accuracy andutility of automatic quality tests used in high-speed production. Amethod is proposed for estimating the probabilities that the production process produces a product which is defective, that thenon-defective product will pass the test, and that a defectiveproduct fails the test. Given these estimates, it becomes possibleto deter mine the consequences to quality and output of using thetest in various ways.

Data for printed circuit manufacture are used to demonstrate themethod. Models of the effectiveness of various product testingprocedures are investigated, the expected net profit is calculatedand the probability of dispatching a defective product to a customerassessed.

For further info - contact Peter Cerone - pc@matilda.vut.edu.au

Young Practitioner's Forum 1996

by
Chair: Dudley Foster
Speakers:Jeremy Howard - ACP, Vincent Yeo - Simsion Bowles Cathy McGurk - ASE

3:30 PM Fri, 20 May 1996
RMIT Building 8, Level 9, Room 66

On Generalized Ekeland's Variational Principle with Vector-Valued Functions

by
Professor Guang-Ya Chen
Institute of Systems Science, Beijing, China

Tuesday 21 May, 4.15 pm, May 1996
Richard Berry Building (Mathematics Dept.), University of Melbourne,Classroom C (in the First Year Learning Centre)

All interested are welcome to come. Direct any enquiries to meon 9344 6761 or craven@mundoe.maths.mu.oz.au . Sadly, we cannot offer you parking at Melbourne University, unless maybe you get a permit from another institution.

On the Relationship between Branch-and-Bound and Dynamic Programming

Dr Emmanual B. Macalalag
Mathematics and Statistics Computing Laboratory
De La Salle University
Manila, Philippines

Friday, 2:15 PM, 31 May 1996
Theatre B, Richard Berry Building
Department of Mathematics
The University of Melbourne

1. their methodological foundations
2. the prominent characteristics of their algorithms, and
3. the factors that affect the efficiency of their algorithms.

We approach our analysis by separating the methodology from the computational aspects of the methods. This has led to a simple yet effective theoretical framework which combines branch-and-bound and dynamic programming together. This framework may also satisfy Marsten and Morin's call for a unifying framework for discrete optimisation.

Solving Large Scale Production Planning Model

David Sier, CSIRO Division of Mathematics and Statistics, and David Noble, Swinburne University of Technology

6:30 PM, Wed, June 19, 1996
Building 8, level 9, Room 66, RMIT

Abstract

In this talk we discuss methods used to try and solve a large scaleMixed Integer Linear Programming Model of a production planning system for one of our clients.

At present, we can solve the problem using reduced data sets but atfull size, the input MPS input data file exceeds 200MB in size and the problem has many millions of variables.

This talk concerns work in progress and we consider different methods for reducing the problem size and the use of heuristics to deal with blending and product separation aspects of the problem.

Problem Definition and Software Design for Rapid Prototyping of OR Solutions

Dudley Foster

Introducing Operational Research based systems to organisationswhich have never used OR before is very different from building an OR system for an organisation which is already "sold" on OR. For external consultants, there is a particular need to deliver benefitsquickly so that the client can sell the new ideas to others in thecompany. All of this puts pressure on the consultant to develop ways of working which meet these needs without compromising thetraditional OR virtues of rigorous problem definition and robustsystems. Workshops have proved a good way of speeding up theproblem definition process and a few simple, software principlesfacilitate the rapid error free development of (spreadsheet based)systems. All of the above will be illustrated with examples fromrecent con sultancy assignments with particular focus on a projectwhich involved the introduction of LP to an organisation which hadnever used it before.

Bilevel Modelling of a Multiperiod Smelter Model

Professor Miles G Nicholls
School of Information Systems
Swinburne University of Technology

4:15PM, August 7, 1996
Classroom C, Richard Berry Building
The University of Melbourne

The NEOS Server for complementarity problems: PATH

ProfessorMichael C. Ferris
Computer Sciences Department, University of Wisconsin -- Madison,1210 West Dayton St., Madison, Wisconsin 53706.

2:15 PM, Monday, August 26, 1996
Thomas Cherry Room, Richard Berry Building, University of Melbourne

We explain the notion of a mixed complementarity problem and how this includes as special cases the problems of square systems of nonlinear equations and the optimality conditions of nonlinear programs. The Network-Enabled Optimization System (NEOS) is the electronic clearing house of the Optimization Technology Center that allows optimization problems to be submitted and scheduled for solution on various local and remote machines using a variety of state of the art solvers. The process of submission of a complementarity problem to NEOS is briefly outlined.

We describe a particular solver, PATH, for the solution of mixed complementarity problems and give details of how this solver is connected to NEOS. These details include a description of how the Jacobian of the defining nonlinear function is derived using ADIFOR, and how the resulting problems are solved on available (idle) workstations at the University of Wisconsin using the CONDOR system.

This represents joint work with Jorge J. More'

Two Talks!!!!!!

Yazid Sharaiha
OR and Systems Group
The Management School Imperial College
53 Prince's Gate London SW7 2PG

2pm-3.15pm, Wednesday 28 August 1996
CSIRO, Division of Mathematics and Statistics Gate 7, No 71, Normamby Road, Clayton 3169

1. A Tabu Search Algorithm for the Capacitated Minimum Spanning Tree problem

   The capacitated minimum spanning tree problem (CSP) is considered. Given a graph G=(V,A), where each vertex v(i) has a traffic load q(i), and each arc (i,j) has a cost c(i,j), the CSP is the problem of finding a minimum cost spanning tree connecting a given root vertex to all other vertices via a set of subtrees, such that each subtree is incident on the root by exactly one arc and has a capacity below a given threshold Q. We consider a tabu searchheuristic for the solution to the problem. The neighbourhood search is based on subtree "cut and paste" strategies. A data structure is proposed to facilitate the on-line updating of the subtrees. Computational results are reported on test problems.

2. A Graph-Theoretic Approach to Line Image Processing

   In this paper, we propose a novel graph-theoretic approach to binary line image analysis. Examples of line images are signatures, Chinese characters and transportation networks. In particular, we consider the thinning or skeletonization problem to illustrate this approach. The image is first mapped onto a graph, where all subsequent operations are to be performed. Analogies between the topological structure of the image and the combinatorial structure of the graph are established. The skeleton-location problem is then decomposed into a series of optimization subproblems on the graph. The problem is essentially formulated as a discrete location problem on the graph. Computational results are presented to illustrate the practical application of this approach.

Yazid Sharaiha received his BSc(Eng) from Imperial College (1986), his MS(Eng) from University of California, Berkeley (1987) and his PhD in Operations Research from Imperial College (1991). He joined the academic faculty of Imperial College Management School in 1991 as a Lecturer in the Operations Research and Systems Group. His research interests include combinatorial optimization and graph theory and their application to image analysis, transportation problems, and newtork design.

Computational Aspects of Mathematical Programs with Equilibrium Constraints

Houyuan Jiang
Department of Mathematics, The University of Melbourne

4:15PM, Wednesday September 4, 1996
Thomas Cherry Room, Richard Berry Building, University of Melbourne

Vehicle Scheduling - Theory and Practice

Jane Parkin
School of Computing and Maths University of HuddersfieldUK

3:30 PM, Friday September 13, 1996
Thomas Cherry Room, Richard Berry Building, University of Melbourne

This seminar provides a brief introduction to the textbook vehicle scheduling problem and discusses some of the most commonly used algorithms used to solve it. It then goes on to discuss the practical requirements of transport planners, giving some details of vehicle scheduling software currently available and case studies of companies summarising their achievements using vehicle scheduling packages. At the end of the seminar, a demonstration of a vehicle scheduling package will be provided for any interested attendees.

Multi Criteria Decision Making; (An overview?)

Fatemeh Ghotb
Swinburne University of Technology

5:30 PM, Wed October 23, 1996
Room D704, Victoria University of Technology (Footscary Campus)

Full information and registration forms available from Paul Lochert or (Dr.) David Dowe, Dept of Computer Science, Monash University, Clayton, Victoria 3168, Australia, dld@bruce.cs.monash.edu.au Fax:+61 3 9905-5146

If you require further information and Registration forms contactConference Action Pty Ltd, PO Box 1231, North Sydney, NSW 2059

Optimization Day 1996

July 18 1996

I have received e'mail from three people who would like to attend, two at least of the three offering papers. However, weneed some more people to present papers, and to attend and listen, inorder to have a viable one-day conference. I am going overseas in earlyJune (will be back for July 18) - so conference organization must bedone before I depart.

Please tell me *quickly*, by e'mail :

• Would you like to give a paper at this Optimization Day? If yes, then tell me an approximate title, so I can advertise it.
• Would you like to attend the Optimization Day, without being committed to presenting a paper?

Yours Optimizingly (optimistically ?),
Bruce Craven
Dr B D Craven, Math Dept, University of Melbourne
craven@mundoe.maths.mu.oz.au
16.5.96

Student Conference

1 - 5 PM, Wed October 23, 1996
Room M101, Victoria University of Technology (Footscary Campus)

Recent Advances in OR

Business Modelling in Excel*

• Wed 13 Nov: Risk analysis, Scenario analysis
  
• Thur 14 Nov: Constrained optimisation, Linear programming
  
• Wed 20 Nov: Statistical modelling, regression analysis
  
• Thur 21 Nov: Forecasting

* Excel is a registered trade mark of Microsoft.

Vehicle Routing/Scheduling/Depot Location

Practical Simulated Annealing

• Chair: Steve Weal
  Department of Mathematics
  Swinburne University of Technology
  P.O.Box 218
  HAWTHORNE VIC 3121
  (W)9728 7131
  (F) 9819 0821
  (E) sew@stan.xx.swin.oz.au

• Vice-Chair: Kaye Marion
  Department of Statistics and Operations Research
  RMIT
  360 Swanston Street
  Melbourne 3000
  (W) 9660 3162
  (F) 9660 2454
  (E)k.marion@rmit.edu.au

• Secretary: Baikunth Nath
  Gippsland School of Computing and Information Technology
  Monash University
  Churchill, Vic 3842
  (W) 9902 6468
  (F) 9902 6842
  (E) b.nath@@fcit.monash.edu.au

• Treasurer: Lutfar Khan
  Victoria University of Technology
  P.O.Box 14428 MMC
  Melbourne 3000
  (W) 9790 4687
  (F)
  (E) Khan@matilda.vut.edu

• Committee:
  
  • Peter Cerone
    Department of COmputer Mathematical Sciences
    Victoria University of Technology
    P.O.Box 14428 MMC
    MELBOURNE 3000
    (W) 9688 4689
    (F) 9687 7632
    (E) pc@matilda.vut.edu.au

  • Paul Lochert
    Department of Mathematics and Sciences
    Monash University - Caulfield Campus
    P.O.Box 197
    CAULFIELD EAST 3145
    (W) 9903 2337
    (F) 9903 2227
    (E) p.lochert@maths.monash.edu.au

  • Cathy McGurk
    A.S.E.
    (W) 9882 7522
    

  • David Noble
    Department of Mathematics
    Swinburne University of Technology
    Edinburgh Road
    LILYDALE 3140
    (W) 9 728 7134
    (F) 9728 7139
    (E) dnoble@banyan.swin.edu.au

  • Moshe Snieodvich
    Department of Mathematics
    University of Melbourne
    PARKVILLE 3052
    (W) 9344 5559
    (F) 9344 4599
    (E) moshe@maths.mu.oz.au
    (WWW) http://www.maths.mu.oz.au/~moshe/

  • Patrick Tobin
    Department of Mathematics
    Swinburne University of Technology
    P.O.Box 218
    HAWTHORNE 3121
    (W) 9819 8013
    (F) 9818 3645
    (E) pct@stan.xx.swin.oz.au

  • Poli Konstantinidis (Student Member)
    Department of Mathematics
    RMIT
    Melbourne 3000
    

• Ex Offcio:
  
  • Santosh Kumar
    Department of Applied Mathematics
    National University of Science & Technology
    P O Box 346, Bulawayo
    ZIMBABWE
    E-Mail: MANGENA@esanet.zw

• OR Solutions: Problem Solving through Quantitative Analysis
  P.O.Box 2086
  1/242 Bambra Road
  Caulfield 3161
  Tel: (03) 9578 6380
  Fax: (03) 9578 2321
  

• Hearne Scientific Software
  
• SiliconGraphics
  Mr Andrew Wyatt
  357 Camberwell Rd
  Camberwell VIC 3124
  Tel: (+613) 9882 8211
  Fax: (+613) 9882 8030

Crawl back to ASOR National
APORS'97