Melbourne Chapter

---

2002 Program

---

Workshop TITLE Dynamic Programming SPEAKER Iwamoto, Maruyama, COle, Sniedovich WHEN 2:00-5:00M, Thursday, Feb 14, 2002 WHERE Theatre B, Richard Berry Blg, The University of Melbourne Program 2:00 - 2:45 PM: Three Dynamic Programmings on Fuzzy Decision-making Seiichi IWAMOTO Department of Economic Engineering Graduate School of Economics, Kyushu University 27 Hakozaki 6--19--1, Higashiku, Fukuoka 812--8581, JAPAN E-mail: iwamoto@en.kyushu-u.ac.jp Abstract: From a viewpoint of stochastic dynamic programming, we develop "Decision-making in a fuzzy environment" in Bellman and Zadeh's seminal paper (1970). It is formulated as a maximization problem of expected value of minimum criterion not over the conventional Markov policy class but over three new broad policy classes. We present three dynamic programming approaches --- (1) membership-parametric method, (2) history-parametric method and (3) multi-stage stochastic decision tree-table method ---, which yield a common optimal policy in a broad class. A complete set of optimal solutions for Bellman and Zadeh's model is illustrated. 2:45 - 3:30 PM: Strong representation of a discrete decision process by a bitone sequential decision process Yukihiro Maruyama Department of General Economics Nagasaki University Katafuchi 4--2--1, Nagasaki 850--8506, JAPAN E-mail: y-maruyama@cc.nagasaki-u.ac.jp Abstract: Making use of automata theory, I will make clear the relation between a given discrete decision process (ddp) and a finite-state sequential decision process (bsdp) which is a sequential decision process (sdp) with bitonicity condition. Assuming that the original discrete optimization problem is given in the form of the process (ddp), I will present a strong representation theorem that provides a necessary and sufficient condition for the existence of the bsdp with the same set of feasible policies and the same cost value for every feasible policy as the given process ddp. 3:30 - 4:15 PM: Generic constructs for sequential decision problems Gavin E Cole University of Melbourne Department of Mathematics and Statistics Email: gecole@ms.unimelb.edu.au Abstract: Modelling sequential decision problems (SDPs) with dynamic programming (DP) has often been considered more an art than science. In this talk an approach to modelling SDPs based on generic constructs is proposed. Generic constructs are present in the mathematical formulation of many problems commonly modelled with DP. They generally model some salient aspect of the problem and provide a direct and obvious link to corresponding components of a relevant DP model. We discuss the links between problem constraints, decomposition, generic constructs and DP models. Lists of generic constructs and problems are presented and a number are discussed in some detail. 4:15-5:00 PM: eDP Moshe Sniedovich University of Melbourne Department of Mathematics and Statistics Email: m.sniedovich@ms.unimelb.edu.au Abstract: In this presentation we shall take a tour of a new project whose goal is to promote the use of DP in teaching, research and practice. The presentation will feature live demonstration of a number of interactive DP modules that are currently available on line via the tutOR and tutORial projects. These modules cover the following topics: Knapsack Problem (0-1, unbounded) Travelling Salesman Problem Shortest path problem (deterministic and stochastic) CPM Tower of Hanoi (min and max, iterative and recursive) Counterfeit Coin Problem (minmax, expected value, other) Prince's Pub Problem Transitive Closure CONTACT Paul Lockert, E-mail: paul.lochert@sci.monash.edu.au

Lecture TITLE Credit Scoring: Development and Methods SPEAKER Jim Marinopoulos, ANZ WHEN 6:00-7:00M, Wednesday, Feb 20, 2002 WHERE Room 8:9:66, RMIT ABSTRACT Credit scoring is a statistical means of providing a quantifiable risk factor for a given customer or applicant. It is a process whereby each piece of information provided in a customers credit application is awarded points which are added together to arrive at a score. The model that determines these points is colloquially called a "Scorecard". Credit scoring was invented in the United States in the late 50s. Although immediately successful, it did not meet with widespread acceptance within the consumer credit industry in the USA until the late 70s and in the UK in the late80s. Credit scoring now accounts for the majority of consumer credit decisions in the USA, UK & Australia and has benefits to both the borrower and lender. This presentation will discuss statistical methods used in credit scoring as well as its use around the world. CONTACT Paul Lockert, E-mail: paul.lochert@sci.monash.edu.au

AGM WHEN 5:30-6:00M, Wednesday, March 20, 2002 WHERE Room 8:9:66, RMIT ABSTRACT AGM -Get your nominations in. 6.00-7.30pm Guest Speaker (see below) CONTACT Paul Lockert, E-mail: paul.lochert@sci.monash.edu.au

Lecture TITLE Professional Ethics and Commercial Imperatives: A Philosopher's Perspective SPEAKER Irena Blonder, Manager, Centre for Applied Philosophy and Public Ethics, University of Melbourne WHEN 6:30-7:30M, Wednesday, March 20, 2002 WHERE Room 8:9:66, RMIT ABSTRACT About the Centre: Irena Blonder is the Manager of the University of Melbourne division of the Centre for Applied Philosophy and Public Ethics (CAPPE). This is one of two divisions, the other one being at Charles Sturt University in Canberra. The Centre promotes community discussion and professional dialogue concerning key ethical problems through workshops, conferences and public lectures and, in its previous form as CPPI, has conducted business ethics consultancies and roundtables with various companies, including Shell, Telstra, BHP, Sydney Water and Western Mining. Irena, who is currently doing research on whistleblowing and organizational accountability, will give an overview of how and why the centre was established and of their work with business and other organisations, such as nurses and the police. This naturally leads on to a common theme in work with professionals: the conflict between two opposing sets of objectives: those created by the aims and standards of their professions, and those of the organization where they work. She observes that professionals often get caught up between two opposing sets of objectives: those created by the aims and standards of their professions, and those of the organization where they work. The former are usually expressed in codes of ethics / practice, or given some other, usually publicly accessible, form. With regard to the latter - most organizations do have a public set of objectives, but they tend to be interpreted in terms of strategies and targets which shift and change, being market- or politically driven. Typically, in a corporate environment, the knowledge and skills of the professional are used to advance the interest of the organization but not of the profession. The professionals are asked, sometimes prevailed upon, to bend the professional standards to assist in the attainment of the desired outcomes. How could such situations be dealt with? Comply? Resign? Blow the whistle? Discussion Structure: After a about twenty minutes presentation, the meeting will be opened up to questions. The committee have undertaken to provide Irena with a number of prepared questions and all members are encouraged to contribute. Questions will also be taken from the floor. Prepared questions should be emailed to Dudley Foster at: dudleyf@ozemail.com.au. CONTACT Paul Lockert, E-mail: paul.lochert@sci.monash.edu.au

Lecture TITLE GRUSS TYPE INEQUALITIES WITH APPLICATIONS FOR THE MOMENTS OF GUESSING MAPPINGS SPEAKER Sever S. Dragomir, Victoria University, Melbourne WHEN 6:00-7:00M, Wednesday, April 17, 2002 WHERE Room 8:9:66, RMIT ABSTRACT In 1994, J.L. Massey [3], considered the problem of guessing the value of realization of random variable X by asking questions of the form : "Is X equal to x?" until the answer is "Yes". Let G(X) denote the number of guesses required by a particular guessing strategy when X=x. Massey observed that E(G(X)), the average number of guesses, is minimized by a guessing strategy that guesses the possible values of X in decreasing order of probability. In this talk, we will show how classical or new Gruss type inequalities for sequences of real numbers [3] may be used to establish bounds for the moments of guessing mappings. This approach complements the methodology developed by Arikan in 1996, [2]. References [1] E. Arikan, An inequality on guessing and its applications to sequential decoding, IEEE Transactions on Information Theory, 42(1996), 99-105. [2] S.S. Dragomir and R.P. Agarwal, Some inequalities and their applications for estimating the moments of guessing mappings, Math. & Comp. Modeling, 34(2001), 441-468. [3] J.L. Massey, Guessing and entropy, Proc. 1994, IEEE Int. Symp. on Inf. Th., Trodheim, Norway, 1994, p. 204. See you Wednesday at the committee meeting. CONTACT Paul Lockert, E-mail: paul.lochert@sci.monash.edu.au

Lecture TITLE INTELLIGENT SIMULATION WITH APPLICATIONS IN FINANCE AND TELECOMMUNICATIONS SPEAKER Felisa J. Vazquez-Abad, EEE, University of Melbourne, Melbourne WHEN 6:00-7:00PM, Tuesday, May 14, 2002 WHERE Room 8:9:66, RMIT ABSTRACT We deal with systems where performance is hard to evaluate in closed form and simulation is used instead, such as pricing exotic financial derivatives or estimating blocking probabilities in a telecommunication network. The case of interest falls within the realm of rare event estimation, because important events happen very infrequently. This means that prohibitively long simulations must be performed in order to achieve a reasonable accuracy in the estimation. To solve the problem it is customary to apply Importance Sampling. Large deviations theory can sometimes be used to prove that variance reduction is optimal under a change of measure known as exponential twisting. Alas! In our case the common methods fail because our random variables have heavy tails and exponential twisting cannot be applied. Instead we propose a parametrized family of changes of measure and estimate the gradient of the variance with respect to the parameters to adjust these through a stochastic version of the gradient method. Examples will include an application in finance and one in mobile telephone networks. CONTACT Gavin Cole, E-mail: gecole@ms.unimelb.edu.au

Lecture TITLE The Illusion of Risk Effects in Stochastic Dynamic Modelling: How to Really Include Them SPEAKER Robin Pope, Centre for Health Program Evaluation, Business & Economics Faculty, Monash University WHEN 6:00-7:00PM, Wednesday, June 19, 2002 WHERE Room 8:9:66, RMIT ABSTRACT OR (Adaptive) Stochastic Modelling, weighting alternative outcomes by (progressively updated) probabilities of each outcome can bankrupt organisations. It ignores the knowledge-based cause-effect chains distinctive to risk (those caused by change in knowledge about which outcome will occur from pre-outcome periods when probabilities of possible outcomes are non-degenerate, to the post-outcome period when these are degenerate - 0 except for the actual outcome by then known to be 1). The omitted knowledge-based cause effect chains involve finance, planning problems, neurophysiological responses (affecting health, productivity). A different periodisation will be explained to recognise, include these ignored effects. CONTACT Gavin Cole, E-mail: gecole@ms.unimelb.edu.au

---

EXECUTIVE COMMITTEE 2001/02

---

Chairperson:	Moshe Sniedovich (Dr) Department of Mathematics and Statistics University of Melbourne Parkville 3052	E-mail: m.sniedovich@ms.unimelb.edu.au Phone: (w) +613 9344 5559 Fax: +613 9344 4599
Vice Chairperson:	Gavin Cole Department of Mathematics and Statistics University of Melbourne Parkville 3052	E-mail: gecole@ms.unimelb.edu.au Phone: +613 9885-1249
Secretary:	Patrick Tobin (Mr) School of Mathematical Sciences Swinburne University of Technology P O Box 218 HAWTHORN 3121	E-mail: ptobin@swin.edu.au Phone: (w) +613 9214 8013 Fax: +613 9819 0821
Treasurer:	Paul Lochert (Assoc Prof) Department of Mathematics Monash University P.O. Box 197 CAULFIELD EAST 3145	E-mail: P.Lochert@sci.monash.edu.au Phone: (w) +613 9903 2647 Fax: +613 9903 2227
Committee:	Lutfar Khan (Dr) Department of Computer and Mathematical Sciences Victoria University of Technology P O Box 14428, MCMC MELBOURNE 8001	E-mail: Khan@matilda.vut.edu.au Phone: (w) +613 9688 4687 Fax: +613 9688 4050
	Natashia Boland (Dr) Department of Mathematics and Statistics The University of Melbourne Parkville VIC 3052	E-mail: n.boland@ms.unimelb.edu.au Phone: (w) +613 8344 5547 Fax: +613 8344 4599
	Dudley Foster (Mr) 23 Wolseley Crescent BLACKBURN 3130	E-mail: dudley@ozemail.com.au Phone: (w) +613 9894 0355 Fax: +613 9894 0244 Mobile: 0417 342 272
	Santosh Kumar (Dr) Department of Mathematics and Statistics The University of Melbourne Parkville VIC 3052	E-mail: s.kumar@ms.unimelb.edu.au
	Kaye E. Marion (Ms) Department of Statistics & OR RMIT 360 Swanston Street MELBOURNE 3000	E-mail: K.Marion@rmit.edu.au Phone: (w) +613 9925 3162 Fax: +613 9925 2454
	Leonid Churilov (Dr) School of Business Systems Monash University CLAYTON 3168	E-mail: Leonid.Churilov@fcit.monash.edu.au Phone: (w) +613 9905 5802 Fax: +613 9905 5159
Student Representative:	Liam Merlot Department of Mathematics and Statistics University of Melbourne Parkville VIC 3052	E-mail: merlot@ms.unimelb.edu.au
Editor:	Harry Gielewski	E-mail: harrygie@ozemail.com.au
Office Manager:	Kaye E. Marion (Ms) Department of Statistics & OR RMIT 360 Swanston Street MELBOURNE 3000	E-mail: K.Marion@rmit.edu.au Phone: (w) +613 9925 3162 Fax: +613 9925 2454
Ex-Officio::	Harry Gielewski (Mr) 28 Kennedy Street RESERVOIR 3073	E-mail: harrygie@ozemail.com.au Phone: (w) +613 9350 4726 Mobile Phone: 0414 650 110