Volume 21           Number 1               March 2002
Published by: The Australian Society for Operations Research Inc.

• Editorial
• The Art of Designing Financial Models ...... Dudley N. Foster
• OR/MS News
• Job Opportunity
• Call for Papers
• Forthcoming Conferences

Happy New Year . Welcome to the trial electronic version of ASOR Bulletin.

This is the first issue of ASOR Bulletin in the year 2002. In this issue, a technical paper on ‘The Art of Designing Financial Models’ is published in the non-refereed article section. From  this issue the refereed papers published will be clearly marked as ‘Refereed’.

ISSN 1446-6678 has been issued for our ASOR Bulletin (hard copy) and may be quoted if required from this issue. We are in the process of obtaining an ISSN for the electronic version of ASOR Bulletin.

In response to our request in December 2001 issue, only a few members expressed their interests to receive the electronic version of ASOR Bulletin. If you are interested to receive only electronic version, you can still send your response to the editor.

There is always a lack of articles being submitted for publication in the ASOR Bulletin. Like any other previous issues, I like to request again all the ASOR members, ASOR Bulletin readers and OR organizations in the country to contribute to the ASOR Bulletin. The guide for contributors is available either from the editor or from the second back page of the Bulletin.

Address for sending contribution to ASOR Bulletin:

Ruhul A Sarker
Editor, ASOR Bulletin
School of Computer Science
Australian Defence Force Academy
Northcott Drive, Canberra 2600
Email: ruhul@cs.adfa.edu.au
 
 
 

Non-refereed Article

Abstract

Based on a review of the author’s own experience, a number of principles are proposed for designing financial models.  The focus is on models used in the evaluation of investment decisions, but the aim is to devise model design principles of general application.   Three principles are identified as crucial: well-planned and consistent layout; separation of input, calculations and results; and designing for sensitivities.  Turning to the design of generic models, two basic designs are put forward for models used for the evaluation of investment decisions, each tailored to the different sorts of flexibility required in different circumstances.

Introduction

Many financial models are conceived in a rush and then ‘grow like Topsy’, with all sorts of unintended and undesirable consequences.  It’s not hard to work out why this happens, but it’s not that difficult to avoid the problem - and the benefits of a disciplined, systematic approach are substantial.

This article focuses on models used for the evaluation of investment decisions, but most of the issues canvassed and model design principles proposed are of a more general application. The discussion starts with one-off‘ models (developed to assist in a single decision) and then goes onto examine some specific issues relating to generic models (used for all decisions of a particular type).

Why the Initial Rush

Many financial models are spreadsheet applications and these are often originally developed for the evaluation of a one-off decision.  In such cases, there is likely to be pressure to get an answer quickly.  Thus, the person building the model may be tempted to concentrate on providing the required numbers, say the NPV or IRR of a project without worrying about the design of the model.  This is perfectly understandable given the situation described - and, in the short term, it is a response which will please the boss.  But what of the longer term - next week instead of tomorrow?

Why It Is Better to Hasten Slowly

Even with one-off decisions, it is unlikely that the initial answer will be the definitive answer.  Things change: the structure of the project gets revised and/or elaborated; estimates of key inputs to the project are refined; senior management challenge assumptions made by technical specialists; negotiations with third parties take an unexpected turn...  The list goes on - and, even if key assumptions and estimates remain unchanged, there is always a need to perform sensitivities: to investigate the effect of changes in key variables (‘what if’ analysis).

If a model has been developed in a rush, without prior consideration to what changes may be required to assumptions, then performing sensitivities is likely to involve ad hoc changes to the model structure as well as changes to the input data.   In a spreadsheet application, part of the problem is likely to be that there is no proper distinction between ‘input’ and ‘model’ (calculations) within the spreadsheet.  If this is the case, changes can be time consuming and the ad hoc development process is error prone - and difficult to audit.

Moreover, if there are several rounds of revisions to data and new sets of sensitivities are requested, the time spent on modifying the model will soon more than off-set any time saved in the initial rush.

A Seminal Example

A good illustration of the above points is provided by a project in which a financial model was required to evaluate the introduction of a new subscription based service.

Among the problems with the initial model design were the following:

· The basic structure of the model had time running horizontally, but the detailed modelling of the main income stream had time running vertically - and in a separate workbook.
· There was no input area as such and there were input cells located at various stages in the calculations – and they were not even clearly labelled as input cells.
· The modelling of growth patterns in the main revenue stream was mathematically sound, but the input was not at all user-friendly for a commercial person doing sensitivities.

We decided to fix these problems, although the work had to be carried out against tight deadlines.  The changes paid off because of the ability to produce reliable results quickly once definitive data did become available.  These benefits of speed and accuracy were equally important when performing sensitivities.

The above combination of circumstances (time to work on the structure of the model, but with definitive data coming in a rush at the end) is not uncommon.  Indeed, it is very likely when there is a multi-disciplinary team working on the design and evaluation of something completely new.  Typically, different groups will be working in parallel.  For this project, the groups included communications engineers, database designers and marketers.  Everyone was working very late (on many days, takeaways were ordered at 7.30 pm and people left the office at 11.00 pm) and no group was able to pass on definitive quantitative information to other groups until the last day or so.  However, people could work effectively in parallel because they were able to exchange structural information at an earlier stage.

Lessons from the Example

Can we summarise the lessons from this example?  And, if so, how general are they?

Looking at the problems noted earlier, we can list the following lessons:
 

(1) Keep the basic layout as simple as possible, by ensuring, for instance, that time runs horizontally throughout the model.  [Time can run vertically if you prefer, although horizontally is generally better – but the essential thing is consistency.  Always avoid having time running horizontally in one part of the model and vertically in another part.]

(2) Have separate areas (in large models separate sheets) for input, calculations and results.  [This basic software design principle is easy to implement in a spreadsheet, but often neglected – even by IT professionals who rigorously observe it in other situations.]

(3) Think of the sensitivities you may want to carry out and parameterise the input accordingly.  [This rule is very important and very general - and worth illustrating with another example.]

(4) To design a model you need structure, not data.  And, if you ask the right questions, people can often tell you about structure before they can provide any useful quantitative input.  [This rule is also very general - and is the key to designing good generic models.]

Another Example with Simple but Powerful Lessons

In this example the model design issues pertain to the modelling of working capital.  The business context was a depot supplying fuels and lubricants to farmers and other regional businesses, but the lessons regarding the modelling of working capital are quite general.

The underlying design of the model was based on the idea of modelling the business in terms of volumes and margins.  The rationale was that it is easier to predict the future behaviour of margins than of the prices and costs from which the margins are calculated.  However, we also realised that working capital cannot be modelled without assumptions about prices and costs, but we wanted to keep this part of the model as simple as possible.  Accordingly, we decided to get round the problem by using the variable Gross Margin as Percent of Sales to derive an implicit price level from the input on margins.

This was not a wise move.  When it came to performing sensitivities, this aspect of the model design proved quite unsatisfactory.  Readers might like to consider why before reading on.

Fundamentals of Design for Sensitivities

Recall the third of our lessons from the previous example: think of the sensitivities you may want to carry out and parameterise the input accordingly.  That meant the model should have separate input parameters for each of:

· Sales Volume
· Price
· Payment Terms

How well did our ‘simple’ approach meet these requirements?  Not very well at all.  The way working capital was modelled, using the variable Gross Margin as Percent of Sales, meant that sensitivities could only be readily performed for one of the listed items, Sales Volume.  For Price and for Payment Terms, the assumptions were both ‘hidden’ and ‘hard coded’.  This is the direct opposite of the recommended approach.  To enable proper sensitivity analysis on Price or Payment Terms, the model had to be completely restructured.  In fact, to provide full flexibility, separate input rows were created for both of these items.

Once we realised what was needed, it was not hard to design the requisite changes.  The key to success is thinking about what is required before considering how to do it.

Designing Generic Models

When we turn our attention to generic models, the consequences of poor design are even more serious than with models used for the evaluation of a one-off decision.  Generic models will often be used by a wide population of staff for multiple decisions over several years.

Obviously, all of the principles discussed so far also apply in the case of generic models.  But, are there additional issues to manage?  Yes there are.  A lot of thought has to be given to the question of how to provide for the full range of different types of income stream and cost.  Further, this has to be done in a way that makes practical sense to all the users and does not compromise central control on the quality and consistency of evaluations performed in different locations.

In the author’s experience two basic designs work well - provided the chosen design is well aligned to the circumstances.  These designs have been labelled the Rigid A La Carte Design and the Flexible Fixed Design.  The two designs are intended for very different situations.

The Rigid A La Carte Design

This model design is ideal for situations where a central management group can tightly specify the different possible types of income stream and cost in advance.  Also, the list needs to be comprehensive.

Is this a realistic scenario?  Yes, it is for things like a franchised business and other retail networks.  The design works well because the model has input areas for all the possible income streams that the user may want to incorporate.  On the other hand, all the things determined centrally can be ‘hard coded’.  For instance, tax depreciation can be ‘hard coded’ and linked to a predetermined list of equipment types.

The results section of the model is easy to build because there will be ‘null data’ in all the sections not used in a particular case.  You can add as many zeros as you like to a result without changing that result.

However, the Rigid A La Carte Design is not suitable for efficiency improving investments in manufacturing or IT.  For situations like that we need the Flexible Fixed Design.

The Flexible Fixed Design

This Flexible Fixed Design is radically different from the Rigid A La Carte Design.  Instead of a long list of tightly specified possibilities, this design works with a much smaller list of income and cost streams, but with a lot of flexibility in the profile of each stream.  Similarly, the input for capital expenditure can be completely data driven.  For instance, tax depreciation rates would be input items because there would be no predetermined list of equipment types.  Thus, tax depreciation rates cannot be ‘hard coded’.

Concluding Remarks

The benefits of good design for financial models should now be clear.  One general benefit of good design will be the elimination of avoidable errors.  Two of the keys to this are:
 

· Keep the basic layout as simple as possible, by ensuring, for instance, that time runs horizontally throughout the model.
· Have separate areas for input, calculations and results.

Next, think in advance of the sensitivities you may want to carry out and parameterise the input accordingly.  Think about what is required before considering how to do it.

Finally, it is worth investing time and effort in good generic models, but this conclusion needs to be tempered with pragmatism.  The Rigid A La Carte Design has been found to work well for investments in a retail network where it is possible to tightly specify the different possible types of income and cost streams.  However, for IT projects and manufacturing investments, it makes more sense to use a shorter list of model components with greater flexibility in each.

For further reading : D. Foster (2002) Course Notes on 'Financial Modelling in Excel', NORCA Consulting Pty Ltd, Melbourne.

Complex Systems

Australian government has just announced a major re-orientation of ARC research funding, with 33% of funds directed to just four areas, of which complex systems is one of them. Operations Research, by nature, fits well with the priority area ‘complex systems’. The detailed announcement can be found in the web: http://www.dest.gov.au/ministers/nelson/Jan02/n06_290102.htm. If you want to take advantage of this, do not forget to apply for ARC grants this year. The details of ARC applications can be found in the web: http://www.arc.gov.au/ncgp/default.htm.
 

MISG 2002

The Centre for Industrial and Applicable Mathematics, University of South Australia organised the Mathematics-in-Industry Study Group Workshop 2002 from February 11 to 15 at their City East Campus. It was another well-organized and successful workshop. Out of six industrial projects short-listed for the workshop, two were clearly from the Operations Research stream. They are:
 

· Scheduling the Charging of Batteries, Exide Technologies,  Peachy Road, Elizabeth West, South Australia, Industry contact: Mr Peter Larner, Moderators: Dr. David Sier, CSIRO   & Assoc. Prof. Graham Mills, CSIRO.

· Identification of Future ADF Vehicles and Trailer Fleets for project Overlander, Defence Science and Technology Organisation, Salisbury, South Australia, Industry contact: Dr. Joanne Nicholson, LOD, DSTO, Canberra, Moderators: Dr. Ruhul Sarker, UNSW@ADFA   & Dr. Hussein Abbass, UNSW@ADFA.

The details of the projects can be found in the web: http://www.unisa.edu.au/MISG/projects.html
 

IFORS On-Line Encyclopedia

On behalf of the International Federation of Operational Research Societies (IFORS) I would like to invite you to a preview of IFORS On-Line Encyclopedia (IOE).

This new initiative will be launched officially at the next IFORS Triennial Conference (July 8-12, 2002, Edinburgh, UK) but the IOE web site is already open for public preview at http://www.ifors.org

Your comments/suggestions on this project will be greatly appreciated.

Moshe Sniedovich, IOE Editor
 

Recent Advances in OR

Australian society for Operations Research, Melbourne chapter organized a one day conference on 13 November 2001 on Recent Advances in OR. The following papers were presented in the conference.
 

· Field Service Territories, Jim Youngman, Institute of Transport Studies, Monash University.

· Forecasting private health insurance premiums for 2002- a soar point? Alan Brown.

· Tax Consolidation, Tom Sandeman, Applied Decision Analysis Group, PricewaterhouseCoopers, Melbourne.

· Recreational Dynamic Programming, Moshe Sniedovich, Department of Mathematics and Statistics, The University of Melbourne.

· Strategy of TORSO RURAL GAME in Southern Africa and its application in developing OR methodology, Santosh Kumar, School of Communications and Informatics, Victoria University.

· The slab loss problem in the paper industry, Robert E. Johnston  and Enes Sadinlija, Australian Pulp and Paper Institute, Department of Chemical Engineering, Monash University.

· Soft and Hard Constraints, Bruce Craven, University of Melbourne.

· Modelling the operation of multi-reservoir systems using decomposition and stochastic dynamic programming, T W Archibald, University of Edinburgh; K I M McKinnon, University of Edniburgh and L C Thomas, University of Southampton.
 

Kaye Marion Receives Francis Ormond Medal

What is Francis Ormond Medal?

The Francis Ormond medal is named after the founder of RMIT who was born in Scotland in 1827 and came to the Melbourne area in 1842.  At the age of 16 Francis Ormond was the manager of a small sheep station owned by his father and by the time he was in his thirties he was a wealthy man.

Francis Ormond had an enormous interest in cultures and educational institutions which broke down the barriers between the rich and the poor.  In 1881, on deciding that Melbourne was lacking in art, science and technological schools as well as working men’s colleges, he offered £5,000 towards the establishment of a working men’s college if the general public was prepared to contribute a like sum.  With the help of the Melbourne Trades Hall Council, which levied its member unions, the money was raised.  The first students walked through the doors of the Working Men’s College, 124 La Trobe Street in 1887.

The prominent grazier and philanthropist predicted that if the college proved as successful as similar institutions he had seen overseas, some two or three hundred students would be attracted in its first two years.  In fact, by the end of 1887, nine hundred individual students had enrolled and by 1891 enrolments were running at about two thousand per term.

114 years after Francis Ormond founded the working men’s college, RMIT University with almost 3 and a half thousand staff and around 55,000 students, continues the tradition of practical, accessible and innovative education and research.

In 1983 RMIT instituted the Francis Ormond medal to acknowledge the importance of our founder and to recognise distinguished service by two members of staff – one general staff member and one academic staff member.

Who is Kaye Marion?

Kaye Marion, the national Treasure of ASOR, was awarded Francis Ormond Medal for 2001. Kaye’s contributions to RMIT, as cited by the Acting Vice Chancellor John Jackson in the awards conferring ceremony, are presented below.

Kaye joined RMIT as a lecturer in the Department of Mathematics in 1973.  She brought to the department a lively approach to teaching mathematical econometrics set in a clear understanding of the business environment honed from a masters in business administration from Monash University. That business approach was refreshing in a mathematically formal department.   It made Kaye a beacon for students keen on experiencing RMIT’s pride in real world applications.   She has supervised more that twenty honours and minor theses.   Her mathematics sat at the interface with the commercial world.   So that when the Department of Statistics & Operations Research formed in 1995, Kaye joined the new department in the role of Director of Consulting.   In this, she helped many RMIT staff and students through the apparent conundrum of interpreting statistical analyses.   Many commercial clients experienced her careful attention to detail and her reports with uncluttered diagrams and clear presentation.

In recent years, Kaye has managed and taught most of the operations research courses in the department.   Her students report on subjects brimming with material and innovative web-based deliveries.   Her leadership in operations research has been nationally significant, having chaired the Melbourne Chapter of the Australian Society for Operations Research (asor) in 1987-1990, and, since 1988 has been treasurer of the national body of asor.

Since 1989 Kaye has been responsible for uplifting RMIT’s spirits as unpaid impresario, promoter and organiser of the ‘Young Performers at RMIT’ Friday classical concerts, each month in the Kaleide theatre.   In this Kaye has provided a platform to enthusiastic RMIT and general public audience for young voice and instrumental students from across the state.   This dedicated support of students and the creation of the much-appreciated musical environment for RMIT and its staff are selfless qualities that lie at the heart of this award.

Congratulations Kaye!
 

Saul I. Gass: A Well-known OR Specialist

In 1975 T.C. Koopmans and L.V. Kantorovich received Nobel Prize for their work in the "Theory of Optimum Allocation of Resources" ………………..

Saul I. Gass

Saul I. Gass is the author of a text on "Linear Programming", now in its fifth edition, has introduced LP and OR to literally thousands of students since it was first published in 1958. He also co-edited (along with the late Carl Harris) the "Encyclopedia of Operations Research and Management Sciences."

Saul I. Gass spent 25 years on the practice side highlighted by stints as manager of the Project Mercury Man-in-Space Program and manager of IBM's Federal Civil Programs, and 25 more years as a highly decorated professor at the University of Maryland.

Gass served as president of the ORSA and Omega Rho (the international operations research honor society) and as vice president for international activities of INFORMS. He's a recipient of the Kimball Medal for service to the society and the profession, the INFORMS Expository Writing Award and the Military Operations Research Society's Jacinto Steinhardt Memorial Award for outstanding contributions to military operations research. Gass was named a Fulbright Research Scholar, a University of Maryland Distinguished Scholar-Teacher and the Dean's Lifetime Achievement Professor at the Robert H. Smith School of Business at the University of Maryland. We could go on and on, but, like we said, it would be easier to list the few things Saul hasn't won, like the Super Bowl. (He did, however, create and host the Knowledge Bowl at INFORMS meetings.)

Gass, 75, retired from the University of Maryland earlier this year, but anyone who knows Saul knows he is not the retiring type. Gass, who used to organize and compete in 10-K runs at INFORMS conferences, just keeps going and going and going, like an Energizer Bunny minus the drum, pink suit and floppy ears. He continues to serve INFORMS in a variety of capacities; he's working on the sixth edition of his text, "Linear Programming"; he's still actively involved in his multiple research interests (LP, large-scale systems, model validation and evaluation, game theory, multi-objective decision analysis and the application of operations research methodologies).

An Interview with Saul I. Gass

An edited version of Saul Gass’ interview with Peter Horner, Editor, OR/MS Today, during INFORMS meeting in Miami (mid-October 2001), is presented below for ASOR readers.

Q: How did you become involved in operations research?

A: I earned my master's in mathematics way back when there weren't any jobs. I applied to various places and sent out 100 letters. The only bite I got was a civil service job in the Los Angeles area. This was in 1949. My wife's family had moved to California from Boston, so I said why not? My B.S. was in education and I thought I was going to teach high school math, but a semester of student teaching in the best high school in Boston convinced me that that wasn't for me.

So I accepted the civil service job. I thought I was going to work at an Air Force bombing range out in the Mojave Desert. It turns out there was a group in Los Angeles called the Aberdeen Bombing Mission doing ballistic work for the Air Force. So I worked in L.A. I wasn't too happy about L.A. or bomb ballistic work, so a few years later I applied for another civil service job in Washington, D.C. They asked me to come out for an interview. I asked if they were going to pay for my trip. They said no. I figured that was the end of it.

 A few weeks later I got a notice to report for duty. That took me by surprise, but I looked into it and the job offer was legitimate. Trudy (my wife) and I and our six-month-old son drove east in my first car, and I joined an Air Force group called the Directorate of Management Analysis at the Pentagon. I had no idea what it was. George Dantzig was the group's chief mathematician. The second day I was there, George gave me some of his papers on linear programming. That's how I got involved in operations research.

Q: Had you ever heard of operations research or linear programming before you met Dantzig?

A: No. But you have to remember this was 1952. ORSA didn't exist until that year. There really wasn't much published on linear programming. No one was teaching LP in schools. I learned LP at the feet of the people who developed it — George Dantzig, Alex Orden, Walter Jacobs, Julian Holley and others who were at the Pentagon at the time.

Q: What kind of work were you doing?

A: I was assigned to the mathematical formulation branch. We set up U.S. Air Force problems — logistics and deployment problems — to be run on IBM accounting equipment geared up to do computational work. We also had access to the first computer (SEAC) at the National Bureau of Standards. The Air Force and Dantzig's group had sponsored the purchase of the second Univac computer. The first one went to the U.S. Census Bureau. The second one went into the basement of the Pentagon. When the first codes were written — not by me, because I wasn't a coder — we would set up problems, check the code, make sure things were going right. Then we would do some production runs, do some analysis and do some research. That's' how I got into the area of parametric programming. I published my first paper in Operations Research in 1953 with Tom Saaty. I think you'll find many people with my background who got into OR the same way I did — by happenstance. They were mathematicians and they just sort of said, "Hey this operations research is some interesting stuff."

Q: You became involved not only with operations research, but with the professional society, ORSA, the forerunner of INFORMS.

A: Yes. I joined ORSA in 1952, shortly after it was founded. I went to a conference at MIT and heard [George] Kimball and others talk about some of the work they were doing. I found out that there were OR summer courses at places like MIT. I took an operations research course from Joe McCloskey at American University, a course on linear programming from Alan Hoffman, and I learned game theory from Al Tucker and Harold Kuhn. It started to sink in that operations research was the type of mathematics that I enjoyed and that I could do well in.

Q: What did OR do with the math that you found so compelling?

A: Operations research basically said, "Here are some real-world problems." They weren't engineering-type problems like how to build a bridge or how to design a jet engine, problems I wasn't good at. OR was about decision problems. How do you choose among alternatives? It just seemed to be an intriguing look at applied mathematics. And there was some nice theory behind it that I could appreciate and understand. That was important.

Q: Decision-making is such a fundamental human activity ...

A: That's true, but it took me — and I think maybe even the profession — a while to figure out that that's what operations research is all about. People ask me what I do. I'll mention operations research, and then I'll go ahead and say that we look at the art, the science and the practice of decision-making. What we do is bring a basic framework to decision-making in terms of how you evaluate complex decision problems.

Q: For years people have struggled to define operations research. You wrote the "Encyclopedia of Operations Research." Give us, once and for all, the definitive definition of operations research.

A: In the first edition, we tried not to define it. We said for those readers who are interested, here's how other people have defined it, and then we quoted various sources including the definition that ORSA was using. ORSA more or less defined it in terms of determining the efficient use and allocation of limited resources. Today, I say that OR is a scientific approach to decision-making. You're applying mathematical and other techniques to decision problems in all areas — business, industry, government and the military.

Q: Ironically, the vast majority of business and industry leaders have never shown any knowledge of, much less appreciation for, operations research.

 A: Let me put it this way: I think those groups that have used the techniques of operations research appreciate it. Go to the petroleum industry, for example. It's all run using LP and related ideas. Take a look at airlines from the point of view of their scheduling. It's all done with OR techniques. They can't help but appreciate the fact that we have some analytical techniques to really help them be more effective and more efficient. Whether or not they realize where it comes from and the history behind it — that it's really operations research — that's another matter. It's the age-old identification problem. I'm on the INFORMS Public Information Committee and the big question we face is, "How do we brand our products and services?" The phrase the committee uses is, "OR Inside." To me, OR is the hidden ingredient in many things.

Q: Does it frustrate you that after 50 years OR still doesn't have anything close to universal brand recognition?

A: I see it a couple of ways. On the one hand, I'm not frustrated because I see the techniques being used and being used well. That's a big plus. On the other hand, when my profession, OR, doesn't get the credit, that's bothersome. You've got to remember I worked out of Washington my entire adult life, and I'm more tied into governmental activities than industry activities. I've seen it in the Washington area many times. When the energy crunch hit us in the 70s, the government rounded up some great people to address the problem. They set up an intricate and novel energy modeling system that was basically linear programming, but there was never any mention of operations research.

Q: It's the same story today, only this time it's "supply chain management."

 A: Everyone talks about supply chain management, which is basically operations research. Certainly, supply chain management would not be possible without all of the techniques and ideas that OR folks have developed over the last 50 years. We can put it together now because we have powerful computers to collect the data. We know how to manipulate it and analyze it, but you don't see it referred to as operations research. Case in point: I was thumbing through the Oct. 3 issue of Fortune magazine in the doctor's office and there was a spread on supply chain management. The article mentioned one of our OR guys, the editor of one of our journals, the head of a supply chain management center. The article had vestiges of operations research all over the place, but you never saw the actual phrase "operations research" used. That's what we're fighting.

Q: Sounds like the name branding is a tougher dilemma than the Traveling Salesman Problem.

A: Funny you should say that. Let me tell you a Traveling Salesman Problem story. A number of years ago, there was an article in Math Monthly, the membership magazine of the American Mathematical Association. I'm reading this article on combinatorics and they mention the Traveling Salesman Problem. The author explained that computer scientists like to use interesting names like Traveling Salesman to describe their famous problems. I wrote a letter to the editor saying the Traveling Salesman Problem did not come out of computer science, it came out of operations research. I gave a history of the name and the Knapsack Problem from Dantzig's original paper. The letter had to go through a refereeing process, but they finally published it. That incident sums up our problem: People from other fields are usurping our ideas and we're not getting credit for them.

Product Manager – Optimization

The official Job Ad can be found at http://www.informs.org/Jobs/ads/2047.html

Here is a synopsis.  The requirement of 5 years experience is negotiable, especially if you have high-level qualifications in appropriate areas, especially constraint programming.

This function is responsible to strategically manage one or many products of the ILOG Optimization software suite (CPLEX, Solver, Scheduler, Dispatcher, Configurator, OPL, Studio, AMPL).

The successful candidate will be expected to work closely with customers and internal industry marketing, sales, R&D and support to identify market requirements and capitalize on ILOG's product strengths and competitive advantage to address the market needs. The need to use market data to define product positioning and market segmentation is an essential skill for the job.

The person will also be expected to develop and implement launch plans of new product versions, write marketing collateral pieces, make pricing recommendations and prepare and deliver training for the sales channels.

Ideally, the best fit candidate will have at least a Bachelor's degree in Computer Science with five years related product management experience or its equivalent. Candidates are expected to have excellent analytical and business research skills. A good understanding of optimization techniques and software development are a plus. They are self-motivated, team player with strong organizational skills. They are also strong in communicating effectively in English, both verbal and written, to high level and technical audiences.
 

Quantitative Analyst - Westpac Options, Sydney

This role will involve the development of option pricing models and systems to manage option-related risks. Requirements include a PhD in mathematics, statistics, engineering, physics or similar (preferably with good exposure to stochastic processes, numerical methods, PDEs, etc), substantial programming experience (preferably C++ on Windows NT), good communication skills, and an interest in the financial markets. Contact: Maciek Blasikiewicz (mblasikiewicz@westpac.com.au).
 

IFORS site:

The IFORS site has an extensive directory of jobs world-wide, not only on the site itself but also through links to WORMS, INFORMS and ORS. The list is intended for job-seekers as well as for persons who wish to know what kind of jobs are available in OR, eg potential OR students.  Many of the jobs which have been advertised  in the local press and in this newsletter also appear there.  http://www.ifors.org/panorama/index.html

Staff Scheduling and Rostering: Theory and Applications
A special issue of Annals of Operations Research (http://www.baltzer.nl/journalhome.htm/0254-5330).

Aim
The principal aim of this issue is to report recent advances of theory and applications and to provide comprehensive reviews in the area of staff scheduling and rostering.

A Note On Nomenclature
The terms staff scheduling and rostering are both used in the literature to refer to the processes involved in linking staff to duties. Which term is used can depend on the application area or perhaps the country of origin of the authors. It may be possible to draw a fine distinction between the terms -- a roster being a list of people who have been assigned to certain duties, and a staff schedule being a list of actions or tasks to which certain people have been assigned -- however the degree of overlap in meaning is substantial. For the purpose of this call for papers we shall take the terms to be synonymous.

Scope
This special issue will be devoted to recent advances in the area of staff scheduling. We also aim to carry high quality reviews on a few topics related to staff scheduling. Original research papers of theoretical and computational orientation will be greatly appreciated. Articles dealing with real-world practice will be particularly encouraged.

Topics
Possible topics for papers submitted to this special issue include but are not limited to:

· Crew scheduling in airlines, railways, mass transit systems, and buses.
· Nurse, clinician and ancilliary staff rostering in health systems.
· Call centre operator scheduling.
· Rostering in emergency services such as police, ambulance and fire brigade.
· Scheduling of retail staff in department stores, supermarkets, franchise chains, and others.
· Rostering security guards.
· General personnel rostering such as post offices, ground staff at airports, hotel staff, and maintenance staff.
· Days-off, shift, and tour scheduling.
· Stochastic staff scheduling.
· Forecasting of demand for services (such as in call centres).
· Determination of staff requirements.
· Leave planning.
· Software survey (for a number of products).
· The application of different solution techniques such as heuristics, meta-heuristics, column generation, mathematical modelling and simulation to crew scheduling applications.
· Other relevant applications and techniques.

Review Process
All papers will be subject to a thorough and stringent refereeing process, in accordance with the usual high standards of the Annals of Operations Research.

Submission Guidelines
Authors of original manuscripts are invited to submit their papers electronically by email to any or all of the Guest Editors listed below, in either postscript or pdf format. Alternatively, 4 hard copies may be mailed to any one of the Guest Editors. To be considered for this publication, papers must be received by Wednesday 31 July 2002. All papers must be written in English, and should not be simultaneously submitted to any other refereed publication.

Guest Editors
Houyuan Jiang, CSIRO Mathematical and Information Sciences, Building 108, North Road, ANU Campus, Acton, ACT 2601, Australia
Email: Houyuan.Jiang@csiro.au

Mohan Krishnamoorthy and David Sier, CSIRO Mathematical and Information Sciences, Private Bag 10, South Clayton MDC, VIC 3169, Australia
Email: {Mohan.Krishnamoorthy, David.Sier}@csiro.
 

New INFORMS Journal on Decision Analysis

The Decision Analysis Society and INFORMS are pleased to announce the creation of Decision Analysis, a new journal on all aspects of decision analysis.  In November 2000 the INFORMS Board of Directors approved the journal, and in May 2001 they appointed Robert T. Clemen and Don N. Kleinmuntz as co-editors-in-chief.  We anticipate publication of the first issue in 2003 and, to accomplish this ambitious goal, we invite you to submit your decision analysis manuscripts for publication consideration. The information below describes the journal's editorial objectives, intended audience, and review process.  Please send us your best work!

Editorial Objectives
Decision Analysis is dedicated to advancing the theory, application, and teaching of all aspects of decision analysis. The primary focus of the journal is to develop and study operational decision-making methods, drawing on all aspects of decision theory and decision analysis, with the ultimate objective of providing practical guidance for decision makers. As such, the journal aims to bridge the theory and practice of decision analysis, facilitating communication and the exchange of knowledge among decision analysts in academia, business, industry, and government.  Articles will contribute to these goals in many ways, using a wide variety of methods and approaches.  Appropriate topics include the discussion of new or existing algorithms, procedures, or processes for implementing decision analysis to real-world situations; and other topics that further the theory and practice of decision analysis. The journal also publishes articles that review and summarize important topics or advances of interest to decision analysts or that provide original historical, scholarly, or practical perspectives on the field. In addition, the journal encourages articles that support the teaching of best practices, such as state-of-the-art applications, case studies, and tutorial articles on decision analysis methods.

Audience
The Decision Analysis audience includes anyone interested in practical aspects of decision analysis and decision theory.  Because decision theory and decision analysis are highly interdisciplinary, our audience in the research community includes specialists in decision theory, psychology, economics, statistics, forecasting, artificial intelligence, decision support systems, operations research, and management science.  Our audience also includes practitioners of decision analysis who work as consultants or as in-house experts within organizations and those who teach courses in decision analysis or decision making in general.

The Review Process
All submissions to Decision Analysis will be fully reviewed.  Our goal is to provide authors with reviews in less than three months.  Authors can help us to attain this goal by adhering to the submission procedures outlined in the instructions to authors.

Editors and Editorial Board
The co-editors-in-chief are Bob Clemen and Don Kleinmuntz.  If you have questions, please feel free to contact them using the e-mail addresses below.  Please submit your paper following the procedures outlined in the instructions to authors at:
http://da.pubs.informs.org/Authors.htm

Robert T. Clemen, Fuqua School of Business, Duke University, Box 90120, Durham, North Carolina 27708 Email: clemen@mail.duke.edu

Don N. Kleinmuntz, Department of Business Administration (MC-706), University of Illinois at Urbana-Champaign 1206 South Sixth Street, Champaign, Illinois 61820 Email:dnk@uiuc.edu

ISORA'2002

The 4th International Symposium on Operations Research and Its Applications
June 1-4, 2002, Yichang-Chongqing, China

Theme and Scope:
The 4th International Symposium on Operations Research and Its Applications (ISORA’2002), organized by The Operations Research Society of China and The Chinese Academy of Sciences, will be held on the Cruiser “San Guo� along the Yangtze River Three Gorges from Yichang to Chongqing in Central China, from June 1 to 4, 2002. The 4th ISORA represents the first of the ISORA series in the new millennium, and has been given the theme "Operations Research in the Information Age". The goal of the 4th ISORA is to provide an international forum for scientists, researchers, educators, and practitioners to exchange ideas and approaches, to present research findings and state-of-the-art solutions, to share experiences on potentials and limits, and to open new avenues of research and developments, on all issues and topics related to the theories of Operations Research and applications. We are particularly interested in submissions that report on experimental and applied research motivated by real-world problems. Theoretical and applied papers are expected to show convincingly the usefulness and efficiency of optimization algorithms discussed in a practical setting. Typical, but not exclusive, topics of interest are: Optimization theory, Algorithm analysis and design, Applications in industry, service, finance, business and military.

Submissions
Authors who wish to contribute papers and/or would like to give a presentation are requested to submit an extended abstract of not more than two pages by email by Dec 1, 2001. Authors will be notified of acceptance by Jan 1, 2002. Final full paper manuscripts (LaTex) should be submitted to the conference secretary given below by Feb 1, 2002. Each submission should start with the title of the paper, the author's name, affiliation, and e-mail address, and a short abstract summarizing the main results of the paper. This should be followed by a scholarly exposition of the ideas, techniques, and a full description of the results. A clear indication of the motivation and comparison with related work should be presented. The paper should not exceed 10 pages using 11 point or larger font. For the detailed format requirement please refer to http://www.orsc.edu.cn/isora02. The proceedings of the conference will be published by China World Publishing Corporation in the series Lecture Notes in Operations Research, and will be available for distribution at the conference.

All registration fees include a copy of the conference proceedings, welcome reception, conference banquet, local bus transportation and all teas/coffee. The registration fee is US$380. Registration fee for the accompanying guests is US$220. Because of the strictly limited number of cruise rooms, on-site registration is not possible. Early registration is strongly encouraged. Please look at the Registration Form for detailed information on payment arrangements.

For more information or to be placed on the mailing list, please contact the Conference Secretary:  Prof. Degang LIU, Institute of Applied Mathematics, Chinese Academy of Sciences, Beijing 100080, China, Tel: 86-10-62541695, Fax: 86-10-62541989, Email: orsc@amath8.amt.ac.cn

Hawaii International Conference on Statistics
June 5-9, 2002
Sheraton Waikiki Hotel, Waikiki, Honolulu Hawaii, USA
(http://hcstatistics.org/hotel_stats.htm)

Submission Deadline:  January 16, 2002
Co-sponsored by the University of Hawaii - West Oahu; and the College of Tropical Agriculture and Human Resources, University of Hawaii.

Call for papers, abstracts, student papers, case studies, work-in-progress reports, research proposals, poster sessions, research tables, or reports on issues related to teaching.  For more information on the format of submissions see http://hcstatistics.org/cfp_stats.htm

Workshop proposals for the pre-session workshops on June 5, 2002 are invited.  http://hcstatistics.org/workshop_stats.htm  for more information.

All areas of Statistics are invited -Agricultural, Applied Bayesian, Biostatistics, Business, Computational, Computer Simulations, Econometrics, Educational, Epidemiology, Industrial, Management Science, Mathematical, Non-Parametric, Operations Research, Probability, Psychological Measurement, Quantitative Methods, Modeling, Teaching of Statistics, and other areas related to statistics.  For a complete list of suggested areas of statistics see http://hcstatistics.org/cfp_stats.htm

Submissions may be made electronically via e-mail to statistics@hcstatistics.org or mailed.  For more information about submissions see http://hcstatistics.org/cfp_stats.htm

Registration Information:  $390 (U.S. Dollars) includes three breakfasts, two luncheons, mid-morning and afternoon coffee breaks, and admission to sessions. Individuals who wish to assist in organizing a session on a particular topic area or in a language other than English please contact statistics@hcstatistics.org. If you would like your e-mail address removed from this distribution list, please respond to statistics@hcstatistics.org and put remove in the subject heading.

IFORS 2002

The sixteenth triennial conference of the International Federation of Operational Research Societies, hosted by the UK Operational Research Society 8 - 12 July, 2002.
University of Edinburgh in the center of Edinburgh, capital of Scotland
http://www.som.umd.umich.edu/ifors2002

This will be a historic conference - the first IFORS conference of the new millennium and a return by IFORS to the United Kingdom, the country that hosted the first IFORS conference and was the originator of OR as we know it.

The conference will feature an eclectic mix of plenary sessions, and invited and contributed talks from practitioners and academics. We hope to have a substantial practitioner contribution, generating a strong interaction between those who have to address the real problems of industry and commerce and those who have knowledge of the techniques and methodologies of Operational Research.

This conference will be at the cutting edge of research. It will not only discuss some of the problems of the new economy but also use new technology - you will be able to submit a paper and book on the web. Programme details and other information about the conference will also be provided via the web.

Conference Secretariat: Chris Barrett, OR Society, barrett@orsoc.org.uk Tel: 0121 233 9300

Important Dates

· Deadline for Abstracts is 15th December 2001
· Early Registration fee applies until 15th March 2002
· Latest date for registration to guarantee inclusion in the programme 30th April 200

OCA2002
The Second International Conference on Optimization and Control with Applications
August 18-22, 2002
Yellow Mountain International Hotel, Tunxi City, AnHui Province.

Topics:

· Duality
· Financial Optimization
· Global Optimization
· Nonlinear Programming
· Optimal Control
· Software of Optimization and Control
· Stochastic Programming
· Structural Optimization
· Systems of Nonlinear Equations
· Variational Inequalities

Invited Speakers including

· N.U. Ahmed, University of Ottawa, Canada
· John A. Burns, Verginia Tech, USA
· Daizhun Chen, The Institute of System Sciences, CAS, Beijing
· Xiuli Chao, North Carolina State University, USA
· Xiaotie Deng, City University of Hong Kong, Hong Kong
· Masao Fukushima, Kyoto University, Japan
· C.T. Kelley, North Carolina State University, USA
· Masakazu Kojima, Tokyo University of Technology, Japan
· John B. Moore, The Australian National University, Australia
· Panos M. Pardalos, University of Florida, USA
· Danny Ralph, Cambridge University, UK
· Kee Roos, Technology University of Delft, Netherlands
· Alex M. Rubinov, University of Ballarat, Australia
· Michel Thera, Universite de Limoges, France
· Philippe Toint, University of Namur, Belgium
· Shuzhong Zhang, The Chinese University of Hong Kong, Hong Kong

Location and Tour
The Yellow Mountains (Huangshan) is one of the most beautiful places in China and in the world.  For registered participants and accompanying persons, there is a free tour to this most beautiful place. There is an international airport at Tunxi, which is connected by flights with Beijing, Shanghai and Hong Kong.

How to participate
If you wish to contribute a talk in OCA2002, please send your abstract before May 15, 2002, to one of the Program Chairs.   The e-mail addresses of three Program Chairs are:

· Wenyu Sun, <wysun@pine.njnu.edu.cn>
· Shouyang Wang, <sywang@rose.nsfc.gov.cn>
· Xiaoqi Yang, <mayangxq@polyu.edu.hk>

Registration
Registration fee for an overseas participant (including hotel fee on August 17-22 nights, meals and tours during the workshop): US$300.
Deposit: 10% of the registration fee needs to be sent to the following account (for overseas participants): Client: Huainan Institute of Technology Address: Huainan Anhui Province China Bank: Huainan Branch, Bank of China Account Number: 03668408093014 before June 30, 2002.

Homepage Address:
http://www.polyu.edu.hk/~ama/events/conference/OCA2002/an1.html  or http://www.math.vt.edu/people/gao/conference/oca2002.html

Optimisation Conference
CUT, Perth, Australia

A Symposium on Industrial Optimisation and an Optimisation Day conference be held in Perth in September 2002. The tentative dates are:

Industrial Optimisation Symposium:  September 30th – October 2nd, 2002.

Optimisation Day:  October 3rd, 2002.

These meetings would be run through WACEIO (Western Australian Centre of Excellence in Industrial Optimisation ).

ASOR’2003
Sydney, Australia

The 17th National Conference of Australian Society for Operations Research will be held in Sydney, from 7 to 11 July 2003, in conjunction with the 5th International Congress on Industrial and Applied Mathematics (ICIAM 2003). The web site for ICIAM 2003: www.iciam.org