- 20th Anniversary Special Publication

Abstracts | Titles Only - 10th Anniversary Special Publication

Abstracts | Titles Only - Proceedings of the Operations Research Seminar: OR for Managerial Effectiveness (21-22 Apr 1989)

Abstracts | Titles Only - Proceedings of the Operational Research Symposium: OR for Decision Support (11-13 Apr 1985)

Abstracts | Titles Only

Back To ORSS Home Page

Mark Goh

Tang Loon Ching

**Bidding for Certificates of Entitlements: Theory and Practical Strategy**

Winston Koh T. H. & David Lee K. C.

**Intelligent System Frameworks for Service Operations Management**

C. M. Khoong

This paper describes a long term R&D initiative to develop intelligent systems for service operations management applications. The initiative is centered around two core applications frameworks, namely manpower/service management and logistics management. These frameworks address automated decision support needs in the rapidly growing and globalizing service industry. The frameworks represent storehouses of applications domain and technological know-how that can be effectively and efficiently translated into a multiplicity of intelligent systems across a spectrum of applications sectors. We describe the philosophy and architecture behind this initiative, and report on developments to date. We also suggest some grand challenges for intelligent systems in service operations management.

**Keywords:**Human resource/OM interface, logistics/distribution, service operations, technology management.**MOLP-PC: An Integrated Decision Support System Environment for Multiple Objective Linear Optimization**

K. L. Poh, M. A. Quaddus & K. L. Chin

We present an integrated decision support environment for multiple objective linear optimization. The system includes a large number of popular algorithms for solving a class of multiple criteria decision making problem known as multiple objective linear programs. An example problem from the literature is solved using MOLP-PC to provide a tour of the system. Actual screen outputs are described which highlight various features of the system.

**Keywords:**MOLP, software, decision support.**Operations Research in Singapore's Defence Establishment**

Sim Cheng Hwee

Operations research in the Singapore defence establishment first started as a small Operations Analysis (OA) group about 10 years ago. The fledgling group of about half a dozen analysts has since grown in size and scope to a community of about 40 deployed in the Ministry of Defence, the main branches of the armed forces and its support agencies. This paper outlines the evolution of the OR community in Singapore's defence establishments, its current organisational structure and the kind of work it undertakes. Some challenges and areas of current research interest are also highlighted.

**Keywords:**Operations research, military, Singapore.**A Manpower Planning Framework with a Special Focus on Long-Term Planning**

Sim Cheng Hwee, Ng Kok-Chuan, Yee Kah-Chee & Ho Tat-Seng

Manpower is a scarce resource. The effective management of this resource becomes even more important as an organisation matures and has to gear itself towards an age of increasing competition. This paper describes a framework for manpower planning containing 5 key modules. The framework also serves as a basis for integrating the various planning models required for manpower planning. A detailed description of the approach adopted for long-term planning is included.

**Keywords:**Manpower planning, mathematical programming, goal programming, linear programming, decision support system.**The Practice of Yield Management in Airline Industry**

Tang Loon Ching

Yield management is concerned with effective management systems for perishable inventories such that the revenue generated from which is maximized. In view of this, it is also referred to as revenue management. Major applications of yield management are towards airline industries in selling air tickets with a wide range of airfares, hotel and rental car industries with constant supplies of rooms and motor vehicles, broadcasters with limited prime time but advertising duration constant. In this paper, we give an overview of yield management and some of the common characteristics across industries where its principles can be applied. This is followed by describing the booking process of airline seat which motivates the main engine involved in yield (revenue) management, i.e. the seat allocation models. We discuss two of the common state-of-the-art approaches in airline seat allocation, pricing and passenger-mix with leg-based origin-destination control. Finally, some remarks on forecasting models and on current deficiency in practical implementation are presented.

**Keywords:**Revenue management, airline seat allocation, bid price.**A Case Study of an Air Crew Scheduling Problem**

Ong H. L., Chew K. L. & Huang H. C.

This paper presents a case study of a crew scheduling problem for an international airline. The study consists of two parts. The first part concerns the generation of an efficient set of flying duties to cover several hundreds of flight sectors operated by the airline. The objecetive function under consideration is to minimise the maximum daily crew requirement. The second part presents a computerised crew management system for the company to generate crew duty rosters. The objective of the system is to generate effective duty rosters for the crew to cover a given set of flying duties, conforming to the even distribution requirement and taking into considerations other non-flying duties to be performed by the crew. The problem is modelled as a series of assignment problems and solved by the assignment algorithm. The system can cope with dynamic changes in flight duties and is able to construct a speedy roster with supporting statistics for the user.

**Keywords:**Crew scheduling, Singapore.**Management Science and the Singapore Postal Services**

Mark Goh & Sun Jie

This paper employs the use of binary Integer Linear Programming, and scenario analysis to determine the optimal delivery base network for Singapore Post (SP). The computer assisted analysis is to assist SP to plan for an efficient mail delivery network in Singapore for the year 2010 and beyond. The study is performed in three phases. The first phase employs an integer programming model to evaluate the optimal number and location of delivery bases. The second phase entails a scenario analysis. Specifically, it fixes the number of delivery bases and varies the number service range to see, if any, the effect on variable operating cost. The third phase fixes the service range of the delivery bases to obtain the optimal number of bases.

**Keywords:**Facility location, integer programming, set covering, case study.

William Hioe

Bill Liu

Chong Fen Sin

Lau Boon Hwee

M. A. Quaddus

**A Brief History of the Society**

Chew Kim Lin

This year, 1987, is the eleventh anniversary of the birth of the Operational Research Society of Singapore. The Management Committee of the Society decided a year ago to publish a special Tenth Anniversary Publication. I am glad to have this opportunity to recollect and gather for this special publication the major events that marked the progress of the Society during the past decade.

**ORSS Survey 1986**

William Hioe

What type of organisation or firm in Singapore is most likely to use OR? What are the most commonly used OR techniques? Answers to these and many other questions are revealed in a survey conducted by the Operational Research Society of Singapore (ORSS) to guage the extent of OR practice and usage in Singapore.

**Findings of the ORSS Membership Survey**

Ong-Tan Yoke Yin

An ORSS membership survey was conducted in June 1986, the main purpose of which was to update membership information and to identify the needs and expectations of the members.

A questionnaire was designed and sent to all the members and a response rate of 30.3% was achieved.

**OR- Its Application and Prospect**

Tan Thiam Soon

**Coping With Technology**

Chang Zeph Yun

**OR in Some Less Developed Nations and Singapore: Some Food for Thought**

Mohammed Abdur Razzaque

**The Role of Management Simulation in Educating Managers and Industrial Engineers**

Andrew Z. Szendrovits

The primary goal of management simulations is to present data in a well designed information system to explain organizational relationships and the interdependence of various management functions for an exercise in group decision-making. Their objective is to reflect the dynamic forces underlying the management of an on-going concern. Completeness rather than complexity is a crucial prerequisite for a model to be useful in providing participants with enduring insight into the nature and pitfalls of managerial decisions. All major facets of planning and control must be included in the simulation model. Operational decisions should focus on marketing strategy, production and inventory control, along with controlling material, labour and capacity utilization in a cost efficient organization which has limited financial resources. Diagrams of the decision network will give an insight into the modelling requirements for a general management simulation.

**OR in Banking: Past Successes and Future Opportunities**

Frederick J. Ridgway

**Military OR- An Overview**

William Hioe

Operational research or operations research (OR), since its beginnings in World War II, has practically pervaded every field of application, be it industrial, financial or commercial. Much has been published about the civil applications of OR but less about the military applications, for obvious reasons. Citing examples from unclassified sources, this paper presents a general overview of the richness of military OR as it is applied to decision-making in the Army, Navy and Air Force. Measures of effectiveness (MOE) used by the military to evaluate alternative courses of action are also discussed.

**Military OR in the Ministry of Defence**

Chow Kay Cheong

This article gives a broad overview of the kinds of problems facing the military that are amenable to analysis using operations research methodology. It describes the existing organisation in MINDEF for operations analysis, and presents some of the studies that have been conducted by this group. Future plans for incorporating operations research into the MINDEF decision-making structure are indicated.

**Short Term Electricity Demand Modelling**

Goh Thong Ngee, Ong Hoon Liong & Lee Yau On

This paper reviews some results in the modeling and forecasting of electricity demand in Singapore. The studies aim at arriving at a reliable approach to characterising the behaviour of short-term fluctuations in electricity demand so as to facilitate forecasts for power system control purposes. The methods proposed, based on univariate time series analysis of data routinely collected by the Public Utilities Board, have been found to be satisfactory in describing both total and sectorial demands, thus enabling future demands in terms of energy and peak power requirements to be forecast. The forecasting accuracies have been verified to be satisfactory from a system management point of view.

**Analysis of Hospital Blood Bank Operating Policies: A Simulation Approach**

Yeong Wee Yong & Chan Wai Chuen

This paper considers a hospital blood bank inventory control problem. A simulation model based on data collected from a local hospital blood bank is used to assist in the determination of the reorder point for blood inventory and to assess the effects of alternative operating policy changes.

**A Scientific Sales Technique for Sales Prospecting and Monitoring**

Bill Liu

**A Bed Simulation for a General Hospital**

Tay Sen Kwan

This paper describes the design and development of a hospital bed simulation model by the Management Services Department, Ministry of Finance. The model is based on a general government hospital with five specialties. It assists hospital administration staff to evaluate the effectiveness of various bed allocation and patient admission policies.

The model generates patients by specialty, medical condition and class preference. Rules used to assign beds to patients are based on hospital policies and practices. If no beds are found, the model puts the patients on a queue until beds become available. When beds become available, transfers are carried out for patients who were not given their preferred classes initially.

The model was designed and calibrated using data collected over a two-month period. Several performance measures, e.g. bed occupancy and number of transfers, were measured and compared with actual results. This was to ensure that the model correctly reflect actual conditions at the hospital.

(21-22 April 1989)

Mohammed A Razzaque

Tan Thiam Soon

**Structured Modeling for Managerial Effectiveness: A Pedagogical Review**

Hum Sin Hoon

The field of MS/OR has made much progress in analyzing and solving models once they are brought into being. The process of modeling itself, however, has advanced only slowly, remaining more on art rather than a science. Structured Modeling is a currently growing area of research that attemps to redress this imbalance. By providing for a formal mathematical framework and a computer-based environment, Structured Modeling specifically enhances the process of conceiving, representing and manipulating a wide variety of models.

This paper seeks to provide an introductory pedagogical review of this growing area of Structured Modeling. In addition, it will also review the ongoing current research work and highlight the opportunities for further work in this area. The objective of the paper is to propagate Structured Modeling within the MS/OR community in our region, with a view to stimulating research in this growing area, and leading hopefully to the adoption and application of this tool to enhance the effectiveness of our managers.

**From Dead Mileage to Sales Territory Allocation- Teaching MBA Students Operations Research**

Yeong Wee Yong

This paper reports the author's experience in teaching a course in Operations Management in the School of Postgraduate Studies, National University of Singapore. In an attempt to motivate students and to bring reality into classroom, a company-based project was used by the author as part of the course requirement. Besides highlighting the main benefits and problems of this approach, the paper also documents some projects students completed in the course.

**Making Marketing Decisions: Who is Likely to Apply Quantitative Techniques?**

Tan Boon Wan & Chan May Lin

In spite of the growing recognition of the usefulness and importance of quantitative techniques to process information for decision making, the extent of adoption by marketers has been low. So who are the users? This question has been studied within an adoption and diffusion of innovation framework in which the use of quantitative techiques has been treated as innovations. Data were obtained from 103 marketers. They were grouped into the three categories of early, majority and late users of quantitative techniques. It is postulated that the three areas of personal characteristics, perception of quantitative techniques and the organization the marketer works in affect his use of quantitative techniques. From literature review and interviews, a set of questions in each of the three areas was derived and factor analyzed to reveal the major factors. These were then treated as predictor variables in the derivation of discriminant functions to classify the marketers. This analysis revealed the major predictor variables that can help identify users of quantitative techniques.

**Review of Vehicle Replacement Policies in a Large Organisation**

M Logendran

For most vehicles, maintenance costs increase with age. After some time it becomes cheaper to replace them with new vehicles than to continue their maintenance. This paper considers the implications for an organisation with a large fleet of vehicles, of using vehicle costs to control its vehicle replacement policies.

We concentrate on the economic aspects of how an organisation should decide when to replace its vehicles. Sometimes a replacement is made on other grounds, such as changed technical specifications or considerations of reliability or prestige. These must remain the responsibility of managerial policy and are not considered in this report.

The paper first describes the theory of replacement and defines 2 replacement policies - group replacement and individual replacement. This is followed by an application of the theory to a department with a large fleet of vehicles.

**Review of the Central Store in a Statutory Organisation**

Johnny Pang

This paper describes a study that was carried out by the Management Services Department (MSD) for a large statutory organisation which was expanding rapidly. The organisation wanted advice on whether it should expand its Central Store to cope with a substantial increase in maintenance materials that would result from the construction of a very large building and several new facilities. This paper discusses the approach that was taken to deal with the problem and the results that were obtained.

**An Aggregate Approach to Manpower Planning and Production Scheduling at an Electronics Manufacturing Industry**

H. L. Ong & T. N. Goh

This paper discusses the application of some OR techniques in production planning for an electronics manufacturing company. The planning process is divided into two levels: aggregated production, and detailed production. At the aggregated production level, the problem is formulated as a linear programming model which minimizes the inventory and production costs, subject to delivery schedules and production capacity constraints. At the detailed production scheduling level, the problem is modelled as a mixed-integer programming problem which minimizes the production and setup costs. Some actual production data from the electronics manufacturing company are used to access the effectiveness of these models and the results are discussed.

**OR Applications in Defence**

Koh Peng Hong

This article gives an overview of the areas within the defence establishment where OR methodology can be aplied effectively to tackle complex problems and issues that the military faces. A generic case study on the evaluation methods to compare fighter aircraft is presented. Please note that the views expressed here are the author's and should not be taken as the organisation's official views.

**Kim Chuan Refuse Transfer Station - Applying a Simplified Transport Model**

Lim Chuan Poh

This paper explains the application of a simplified transport model in a feasibility study conducted by the Ministry of the Environment (ENV). The 1982 study was conducted to assist ENV to assess the cost-effectiveness of constructing a refuse transfer station. The 1500 tonnes/day Kim Chuan Refuse Transfer Station was constructed in 1985 and has been in full operation since 1986. The study showed that the total project cost of $30 million was recoverable within 2 years of the transfer station operation, through savings from manpower and vehicles.

**Managing Business Information Effectively**

Peter Connell

Strategic Information Value Analysis is a management tool developed by Arthur D. Little to help managers plan their information systems. Since information supply is a critical issue for OR practitioners, SIVA represents a technique which can help them to influence management in improving that supply. The planning approach is top-down, driven by overall business objectives and future information needs, and aims to produce a set of four interlocking information architectures: information supply, data, technology and organisation. The information modelling techniques which SIVA uses can help both managers and OR analysts to manage business information more effectively.

**Crew Rostering in a Service Department**

K. L. Chew & C. P. Low

In this paper, an algorithm for generating crew rosters in a typical service department is presented. Formulation of the algorithm and an illustrative example are included.

**Application of Non-linear Programming to Currency Portfolio**

Hui Tak Kee

By investing in a portfolio of currencies, the risk as against the numeraire currency can be substantially reduced. A set of efficient portfolios is constructed by using the quadratic programming techniques. The objective is then to find those portfolios that provide the best risk-return combination.

**Development of OR Software for Teaching and Research**

H. L. Ong, H. C. Huang, W. Y. Yeong & K. F. Wong

In this workshop, we discuss the design and development of a series of operations research (OR) packages for teaching and research at the National University of Singapore. Even though these packages aim at supplementing the OR courses taught at the University, they are sophisticated enough to serve as OR tools for modelling, solving and analyzing real-world problems. The OR packages are written in FORTRAN 77 and implemented on the IBM 3081 mainframe as well as the IBM PCs. The microcomputer version of OR packages are user-friendly and easy to use. Each package includes a full-screen data editor, on-line help and facilities for viewing the output on screen.

(11-13 April 1985)

Chew Kim Lin

Ong Hoon Liong

Chong Fen Sin

Yap Kim Yew

M. A. Quaddus

Ong-Tan Yoke Yin

**Economic Order Quantities/Economic Review Periods for Multi-Item Stock Control**

Walker, J.

The use of the standard economic order quantity/economic review period (EOQ/ERP) formulae to a very large number of stock items may result in a major change in the total number of orders per annum. In such a case the predicted savings may not be achievable since fixed costs of personnel and/or equipment many need to be disposed or provided. A simple modification to the EOQ/ERP formulae is proposed. The modification is such that the expected total "active" stock capital per annum is minimized subject to the constraint that the expected total number of orders per annum if equal to its present value. The solution will then at least be workable, and generally will lead to savings in comparison with the present policy. If required an assessment of the effect of changing the total numbers of orders per annum can be attempted by a comparison of the change in the expected total order cost per annum with the change in the expected total active stock capital per annum. An incremental approach is suggested.

**Case Study: Optimisation of a Proposed FMS Production Layout Through Simulations**

Yu, W. S. & Ho, N. C.

A multinational electronic company produced several million tuners a year for colour and black and white T.V. sets in its Singapore plant. The tuners were sold worldwide to its subsidiaries as well as other companies that produced television sets. Currently, testing and fixing of the faulty tuners were done manually.

The company planned to set up a FMS testing and fixing facility, automating many of the testing functions to reduce labour cost and increase productivity. A simulation study was condicted to identify the most appropriate configurations, and furthermore to optimise the most preferred configuration to achieve the planned production target.

**A Simple Graphical Method for a Production/Purchasing Problem**

Dyer, M. E. & Walker, J.

Consider a company which is contracted to supply units of an indivisible product over a finite time horizon with no shortages or backordering allowed. Assuming that the inventory holding costs are so small they can be ignored and the production/purchasing cost function is stationary and convex non-decreasing (no knowledge of this function is required) over the time horizon, a simple graphical method ofor determining an optimal production/purchasing schedule is presented. The graphical method readily facilitates sensitivity analysis of the optimal schedule.

**Strategies for Reducing Inventories in Production and Distribution Systems**

Johnson, L. A.

Singificant economic benefits can result from programs to reduce inventory levels in production and distribition systems. A systematic effort to modify or eliminate factors that would otherwise justify inventory is required. These factors are identified and discussed. Approaches to inventory reduction are described. Types of decision analysis are characterized, and examples given. Information needs and analysis modes for planning, economic justification, and control are discussed.

**University Timetable Construction Under Constraints**

Loo, E. H., Goh, T. N. & Ong, H. L.

This paper discusses a computerized timetable scheduling system which uses a heuristic method to satisfy complex conditions and requirements. The constraints of colleges and universities are taken into consideration. These are soft constraints and can be easily modified or logically removed by timetable planners without changing the scheduling algorithm.

**A Man-Machine Interactive Heuristic Algorithm for Timetable Scheduling at a Polytechnic**

Ong-Tan, Y. Y. & Foo, K. S.

This paper discusses a man-machine interactive approach to obtaining quality timetables for lecturers at polytechnic level.

Initially, the time slots for each day are ranked according to lecturers' favoritism, the algorithm then produces time-tables with uniform total rank score for each lecturer. With the method, special requirements are pre-assigned manually. The computer then matches elements in two sets, M and F, where each element of M consists of a string of information on a lecturer, a class/group, and a teaching activity; while each element of F is made up of a string of information on facility and time available. The matching is done so as to ensure that each time-table obtained in solution has about the same total rank score.

The heuristic algorithm also takes care of the constraints. The given list of constraints are weighted according to the order of importance. By means of the method, more than one timetable may be produced for each lecturer with different violation scores to the constraints. It terminates when an acceptance score is reached.

**University Examination Scheduling Problem- A Systematic Approach**

Yeong, W. Y. & Allaudeen s/o S. Hameed

In this paper we present a three-stage model for examination scheduling -- a problem faced by administrators in all educational institutions at the end of each academic year. A heuristic procedure will be proposed for each stage of the model and the application of the model to a real-life examination scheduling problem will also be discussed.

**Goal Programming Model for Manpower Planning in a Textile Mill**

Prasad, S. R. K., Prabhakar, R. & Selvaraj, C. M. K.

**An Interactive System for Power Generation Planning with Multiple Objectives**

Quaddus, M. A. & Goh, T. N.

An interactive system for multiple objective decision making is presented in the context of an application in electric power generation planning. It accomplishes the work of decision support system because of its flexibility, interaction, and on-line features. The system is briefly described and results of application are analyzed.

**Strategic Decision Support Systems Via System Dynamics**

Paul, H.

System dynamics has been acclaimed to provide a strong methodology for corporate modelling and policy analyses. Couple with some optimization and other mathematical routines, this computer simulation technology can be used for developing a model base for a comprehensive strategic decision support system (DSS). This paper discusses how system dynamics provides the decision-maker with an easily-learned modelling technique for analysing strategic decision probelems.

**Development of Scenarios on Flight Profitability**

Lavoie, G. & Petitjean, D.

The objective of this project was to provide an easy-to-use and inexpensive set of analytical tools, which would give the mid-term planner a complete insight into the economic operating characteristics of his carrier's routes. The mechanics of the package that emerged from the migration of the mainframe files to microcomputers can be partitioned into two distinct categories:__transation procedures__on fields and records, and a__simulation scheme__. In the first case, data bases are created or updated; financial indicators are derived using a set of program files; results of the computations are displayed, sorted or massaged in various ways. A further step leads to the development of "What if?" scenarios, and to their assessment by means of simulation programs, in order to assist in the evaluation of alternatives (routes and/or aircraft types).

**Ground Combat Advance Rate Estimates for Decision Support**

Iyer, K. S. S.

Large scale war games or simulations are useful tools in deciding force structure to support different conflicting situations that may arise. Success of war gaming exercises depends largely on predicting ground combat advance rate of the attacker. In this paper, we will discuss recently developed methods to estimate analytically the advance rate. Part I of the paper gives a detailed investigation carried out by RMC Systems Group, USA and Shape Technical Centre, Netherlands. It presents a method for advance rate in terms of interacting parameters between the inputs of attacking and defending forces by using multiple regression analysis. The predicted equation is being tested with the historical data from World War II. A comparatively simpler method developed by Col. Dupuy based on standard rates of advance is presented in the second part. The Combat Power of each side is calculated by applying all combat operational variables with the use of Operational lethality index. The advance rate is estimated using standard advance rates, power ratio and reduction factors for environmental factors. The relative merits and demerits of the above two approaches are analysed and a stochastic approach based on time, environment and resources is proposed for better monitoring of decision support.

**An Algorithm for finding a Minimax Path in a Multi-Parametric Network**

Tung, C. T.

Multi-parametric path problems have not been studied intensively in the specialized literature, despite their potential applications. Many problems concerning the design and the efficient use of transportation, transmission and communication networks can be treated as multi-parametric path problems. In this paper we are concerned with a special class of multi-parametric path problem; that is to determine a path between a specified pair of nodes, which minimizes maximum (f, f_{2}, ..., f_{m}). For instance in a road network several parameters such as time, distance, cost, etc. can be assigned to each arc. After multiplying some weights to the parameters, the decision-maker may be interested in minimizing the worst-case of the objective functions, f, f_{2}, ..., f_{m}. An implicit enumeration algorithm, which is based on the multiple labeling scheme and heuristic decision, is presented. Validation of the algorithm and an illustrative example are included.

**On a Single Depot Multiple Vehicle, Node Routing Problem**

Loh, Y. K. & Yeong, W. Y.

A government department in Singapore is facing with the problem of routing a heterogeneous fleet of vehicles from a central depot to customers with known demand scattering around the whole island. As the set of customer demand points changes from one day to another, an efficient vehicle routing procedure is essential. This article proposes a model for the problem, discusses its implementation, and suggests further research in the area.

**A Crew Scheduling Problem for the Italian Railways Transport System**

Mingozzi, A. & Ricciardelli, S.

In this paper the problem of building and scheduling crew work periods in the Italian railways system is considered. It is shown that this problem is equivalent to finding a minimum cost circuit in a graph with additional constraints. Two different formulations of the problem are proposed: the former is based on Benders decomposition and the latter on dynamic programming. Bounding procedures and branching strategies for the above formulations are, then, discussed.

**A New Dual-Based Heuristic for the Simple Facility Location Problem**

Tcha, D. W., Ro, H. B. & Meoung, Y. S.

The dual-based approach developed by Bilde & Krarup and Erlenkotter independently has been widely accepted as the most powerful tool for solving the simple facility location problem and a variety of its extensions. Though the performance of this approach is generally satisfactory, there still exist many cases where a better solution than the one yielded by this approach is desired even at the expense of marginal increase of computational load. One representative situation would be when in a branch and cound (b&b) solution procedure, one wants to trim off the b&b tree size by having a good solution (bound) at the initial node (candidate problem). This observation motivates us to develop a new dual-based (or parametrized Lagrangian relaxation) heuristic which emphasizes the solution quality rather than the computational load.

In our dual-based heuristic, several features have been incorporated into the efficiency and consistency of the algorithm. Three major ones are: the modification of the generating procedures of both dual solutions and their corresponding primal solutions, and the introduction of a new primal interchange procedure which further improves primal solutions at hand.

The efficiency of the algorithm was tested with some static and dynamic location problems. The computational results were encouraging: for the absolute majority of test problems, the algorithm generated optimal solutions in a satisfactorily small computation time and even for the remaining problems, the quality of solutions generated was superior to that by the existing dual-based solution procedure.

**TRANSMOP: A Model for Transport Planning Decisions**

Varaprasad, N., Chin, H. C. & Loh, C. K.

This paper describes the development and structure of a microcomputer-based model for transport planning. While based on the conventional four-stage process of trip generation, distribution, modal split and assignment, the model includes some enhanced features not found on commercially available software such as separate network descriptions for private and public transport, and calibration facilities. It is also sufficiently general in structure and flexible in use to allow decision-makers to try out potential network configurations for more detailed testing.

**MICROBUS: A Model to Evaluate Bus Network**

Chin, H. C., Varaprasad, N. & Loh, C. K.

This paper describes a micro computer model, MICROBUS which was developed for bus operators to examine the impact of different policies on various network options. MICROBUS models travel behaviour of bus passengers, including details like transfers, multi-routing and fare structure. It outputs important information, such as load factors and revenue collected, which are indicators of the performance of the services provided.

**Development of Decision Support Systems in the Field of Traffic and Transportation Management**

Bielli, M. & Cini, M.

The increasing development of information technologies and new high level programming languages as well as last theoretical contributions in Artificial Intelligence have set new and interesting instances in the field of Decision Support Systems (DSS).

But, although the availability at decreasing costs of many tools, as videographics systems, data-base systems, personal computers and local area networks offer new chances in supporting managers for decision making, a lot of work has to be done in order to implement and build up effective DSS for practical application.

In this work the peculiarities and the basic structure of a DSS are defined and the differences with respect to management information systems, expert systems and management support systems are remarked.

This paper, among the components of DSS, in particular deals with the formalization of decisional problems, the mathematical modelling and the structure of the models-base management system.

In order to come into practical examples, some typical decision problems are illustrated relatively to the centralized management of urban traffic and resources allocation in a transit company.

Finally the data-base and models-base needs relatively to these problems are analyzed for the design of DSS at operative level.

**Multiple Objective Fractional Transportation Problem with Impurities**

Saxena, P. K.

This paper is devoted to the study of multiobjective transportation problem with impurities in the commodity. The objective functions considered are - (1) Total cost of transportation of commodity to be minimal, (2) Total deterioration of commodities during transportation to be minimal, (3) Amount of commodities to be transported to be maximal. The problem having usual transportation restrictions together with capacity and impurity constraints is converted into a synthesis - fractional objective function to get an efficient solution of the original problem. The approach here has been to transform this synthesis fractional transportation problem into a three dimensional transportation problem and then to find a local minimum of the given problem. Using simplex multipliers technique, the optimality conditions for local minimum are derived and an algorithm is developed for finding a local minimum which occurs at an extreme point of the convex set of feasible solution.

**The Use of Capital Budgeting Techniques in Malaysian Companies**

Han, C. K. & Mohamad S.

Since Joel Dean's 1951 poineering text, the subject of capital budgeting has become deeply entrenched in the finance and accounting literature of the West. During the past 30 years, several studies have reported on the capital budgeting practices employed by business enterprises in the developed countries. These studies were mainly designed to determine the extent of usage of capital budgeting techniques by a variety of company classifications. The surveys conducted have ranged from those merely identifying which technique is used to those which go into great detail to explain what discount rates are used, how cash flows are measured, whether multiple cut-off rates are employed, what factors are important in the capital budgeting decision, as well as a host of other significant factors involved in the capital budgeting process.

In general, it is observed that usage of sophisticated capital budgeting techniques has increased among business enterprises in the US, Europe and elsewhere. Discounted cash flow techniques, in particular, are used more and more by large companies. In South East Asia, however, empirical evidence on the state of practice in capital budgeting is almost non-existent. In an attempt to fill the gap in knowledge, a survey was conducted in October 1983 to study the extent of usage of various capital budgeting techniques and the management of capital expenditure activities in Malaysian companies.

The analysis is based on responses from 66 companies listed in the Kuala Lumpur Stock Exchange with significant annual capital expenditures in the years 1981-1982. An important finding of the survey is that payback period, degree of necessity or urgency, accounting rate of rate (ARR), internal rate of return (IRR), net present value (NPV), profitability index (PI), and net future value (NFV) are, in that order, the most frequently used methods of appraising capital investment proposals. Most companies used multiple techniques when evaluating major capital expenditures. Use of discounted cash flow (DCF) methods is a relatively recent phenomenon, with two-thirds of the DCF users adopting the method only within the last five years. The interest rates most commonly used in DCF methods are target yield for new investments, cost of capital plus risk factor, and cost of long term debt plus risk factor. In three-quarters of the companies, formal proposals and profitability analysis are required for most projects. Market prospects, competition and political risk are perceived to be the most significant factors affecting the companies' capital expenditure decisions. Most companies compute cash flows after deducting either interest, or interest and tax, or interest, tax and dividend; whereas, risk is mainly accounted for by shortening the payback period and setting a high IRR for acceptance.

The 1983 survey also looked into various matters relating to the administration of capital expenditure activities. It was observed that nearly all responding companies were undertaking active and systematic search for alternatives to their major investment proposals. Three-quarters of the companies carried out long-range planning, with a three to five years horizon, with respect to their capital expenditure. Nearly 80 percent of the companies have performed a post-completion audit but these exercises had largely been financial reviews conducted by their accounting departments. The capital budgeting activities in about half of the companies are supported by their computer-based information systems. Operations Research has been used as an aid to capital budgeting decision making but the results of its usage are quite uncertain at this point in time. Generally, the respondents ranked industry-specific information to be more relevant and important than general macroeconomic information in their capital budgeting exercises.

As a final analysis, the responding companies are found to fall into three clusters and can be conveniently grouped as those with annual capital expenditures less than M$1 million, between M$1 million and M$10 million, and more than M$10 million. The sophistication of the capital budgeting techniques employed are then analysed with respect to these three groups of companies. The results are also discussed in the context of a three-stage model of the capital budgeting process.

**Practical Benefits of Executive Decision Support Systems**

Owen, D. & Volpato, M. C.

The broadening exposure of DSS activity at the senior executive level calls for a systematic approach which helps identify how, what and when benefits can be realised in a formal justification.

Senior executive decision-making (predominantly discretionary) and strategic Corporate support depends on proper DSS implementation strategies and well defined control processes.

This paper discusses the above delineation of benefits and their impacts in light of an implementation at ICL Asia Pacific Pty Limited Headquarters in Sydney.

Executive views and perspectives on future advantages are cited in response to the increased needs of executive support in leveraging Corporate effectiveness.

**Application of O.R. to the Railway Management in China- Current Situations and Prospects**

Li, Z. Z.

The paper is divided into four sections. The first section gives a brief development of the applications of O.R. to the railway management in China for the recent thirty years. Several research results are presented in the 2nd, 3rd and 4th sections. They are: finding an optimal solution to a special kind of transportation problem in which replacement occurs by LP approach; determining the optimal organization of the wagon flow by DP method; computing pass capacity of a railway station by queueing theory and probability theory. At the end prospects for the applications of O.R. to the railway management in China are given.

**Thoughts on a Stage Theory of Operations Research**

Han, C. K.

The process of adopting Operations Research in computerised environment is structured in terms of a sequence of stages, from the initial financial-oriented techniques based on accounting systems to the advanced stage of fifth generation computers. Use of Operations Research to support various categories of management decision is the primary criterion variable for demarcating the four stages - each with discriminating characteristics in terms of decision problem, techniques used, supporting information technology and location of Operations Research activities. The stages are driven by the development of computer-based information systems, which in turn are made more efficient and effective by Operations Research. Besides suggesting directions for research that have not been previously raised, the stage theory provides a strategic perspective on the successful introduction and implementation of Operations Research in the enterprise of Malaysia and Singapore.

**An Efficient Linear Search Technique for Multivariate Minimization Algorithms**

Phua, K. H. & Ang, C. H.

In many practical studies one constructs a function of several real variables, F(x_{1}, x_{2}, ..., x_{n}) say, and one wishes to obtain the values of the variables that give the least value of F(x_{1}, x_{2}, ..., x_{n}). In this paper, we consider numerical algorithms for calculating the optimal vector of variables in the case that numerical values of the derivatives of the objective function F(x_{1}, x_{2}, ..., x_{n}) are not available to an automatic algorithm. Recently, little attention has been paid to the problem of one variable minimization. This may be due to experiments which indicate that accurate line searches are inefficient. However, most present-day algorithms to extremize the value of a function of several variables rely upon "linear search" which find an extremal value in a specified direction. Since these linear searches amount to extremizing a function of one variable, an efficient method of one-variable function extremization is important. In a recent study, pratical performances of various linear search techniques are studied and evaluated over a wide range of test functions. Some of these results are reported in this paper. Basing on the relative merits of these searching techniques, an efficient linear search algorithm is proposed. Computational results show that the new algorithm is more efficient than some other optimization techniques considered.

**Games of Discrete State Spaces**

Zhang, S. K.

Tsokos [1] has given a summary on some research work in game theory. Generally speaking, most people when studying the strategies of game theory frequently limit their discussion to abstract spaces. But, in view of the recent advances in science and technology, we must discuss the strategies of discrete state spaces. In this paper, we shall discuss two-person-zero-sum games, the strategies of the games are infinite sequences and we shall prove the principles of the dual games together with the relationship of the values of the games.

**A New Continued Fraction Algorithm and Its Convergence**

Chen, K. Z. & Wang, K. M.

This paper proposed a new linear search method, Continued Fraction Algorithm, which used the continued fraction interpolation [1,4]. The method needs only to compute three functional values at each iteration, and to use one division in computing every coefficient of the continued fraction interpolation. The numerical stability of the algorithm is better. Moreover, the convergence rate is fast. The order of convergence is approximately 1.8393. Thus, this algorithm is a very attractive method for the local one-dimensional optimization. If we use this method to the imprecise linear search, it is necessary only to compute three functional values. This method can also be applied to solve the transcendental equation.

**Multicriteria Optimization of Dynamic Decision Process**

Wu, C. P.

Multicriteria optimization theory and its applications has become a subject attracting wide attention and interest of academic and practical researchers. The motivation of its developent is of two aspects, the one is the challenge given by theoretical problems which have arisen in this subject and the other is the demand to solve practical optimization problems which are of multicriteria type in their nature and are widespread in human activities.

In contrast to the optimization problems with scalar criteria, where in most of the cases either the optimal solution is unique or there exist only a few ones, to the optimization problems with vector or other non-scalar criterion, the optimal solutions form a set containing infinite elements in general, provided that all criterion functions are equally set to a certain extent for optimization by increasing (or decreasing) each of their values as great as possible and there exist no additional objective or subjective optimization considerations. This peculiar characteristics of multicriteria optimization problems originates in the fact that the order relation on the set of feasible solutions of the problem is a partial ordering in general while that of scalar criterion optimization problems is a total one. The set of optimal solution for multicriteria optimization problems is known as the set of efficient solutions or Pareto solutions.

In practice, in order to obtain a final optimal solution determined uniquely, additional optimization consideration such as the preference thinking of the decision maker must be taken into account. However, it is evident that the final optimal solution should be an element of the set of Pareto solutions, no matter what additional considerations have been taken into account for choosing it. In this sense the set of Pareto solutions forms a rational and objective basis for the solution of multicriteria optimization problems and thus a most important object of study in this subject.

This paper aims at giving a brief survey of recent results on the multicriteria optimization problems of controlled dynamic systems, which mainly includes those obtained by the author. Owing to the fact mentioned above, the Pareto optimality will be regarded as the underlying concept of multicriteria optimality and the discussions will be confined to characteristics of Pareto optimal solutions and methods of obtaining them.

**Computer Simulation of the Operation of an Equipment Marshalling Area During Mobilisation of Civil Resources**

Looi, C. K.

This paper presents a computer simulation of the operation of an equipment marshalling area during the mobilisation of civil resources. The objective of the simulation is to evaluate the capability of the EMA to cope with its expected workload. We discuss some of the issues involved in the process of developing the simulation model and program, and running simulation experiments on the model.

**Computer Simulation as a Decision-Making Support for Production Systems Optimisation**

Lee, L. C.

A comprehensive model has been developed to simulate closely the events assiciated with the throughflow of jobs in a production system. The model allows for different machine types and numbers, job patterns into the system and routings within the system, and manning types and levels. A number of characteristics such as job scheduling rules batch sizing and inter-operation transit times are examined. System effectiveness recorded include machine utilisation, job throughput time, mean job queueing time, job lateness and work-in-progress level.

Results indicated that scheduling rules have a significant effect on system performance. Contrary to popular belief, small batch sizes have been found to perform well when operated in association with an appropriate priority rule. An increase in job inter-operation time resulted in an accumulation of work-in-prgress but left machine utilisation relatively unaffected.

**The Use of Statistical Software and Simulation System as Decision Support Tools**

Yeo, G. K.

In a decision-making situation, a statistical software and a simulation language may play complementary roles, with the former providing the data analysis power before and after the modeling process and the latter is used as a modeling tool. While a sophisticated statistical package such as SAS may have language elements that allow users to carry out modeling and a general-purpose simulation system such as GPSS also provides facilities to define tables for collecting statistics, each has its own merit in its primary function as the underlying data structures in the two tools are very different. This paper discusses the integration of these two tools in on decision support application.

**Decision Support Systems in the Airline Industry**

Yu, W. S.

It is not commonly realised how complex an airline is. Typically an international size airline has a turn-over of several billion dollars oer annum, and employs more than 10,000 staff. Its diverse activities encompass in-depth mechanical and electrical engineering, finance, marketing, computer services, large scale meal production, etc.

As a result, decision support systems have been widely used in the airline inductry. Coupled with the peculiar nature of the airline business, a separate decision support association was formed: the Airline Group of International Federation of Operations Research Societies (AGIFORS).

This paper describes the decision support systems that have been successfully implemented in the airline industry. Many of these systems can also be used in other industries.

The author worked for Quantas Airways Ltd in Australia for 6 years as Senior O.R. Analyst. In this paper, he borrowed heavily on his experience at the company.

**DSS: A New Buzz-Word or a New Class of System?**

Raman, K. S.

The concept of Decision Support System (DSS) as a new class of system to support managers in ad-hoc, semistructured and unstructured tasks has emerged strongly in recent years. This paper traces the evolution of DSS and examines the characteristics and capabilities which distinguish DSS from the traditional information systems. Further, this paper explores the roles of models, WHAT IF? and GOALSEEKING analysis and OPTIMIZATION in DSS*.

**SPDMU: A Software Package for Decision Making Under Uncertainty**

Gupta, O. K. & Agrawal, S. K.

A decision maker is often faced with the problem of making a decision under uncertainty. The situation gets complicated because often he has to select precisely one alternative out of many available under several possible uncertainty events out of which exactly one will actually realize. Various quantitative approaches have been developed to deal with such situation. This paper introduces a computer-aided-interactive learning program SPDMU for decision making under uncertainty. SPDMU is interactive in the sense that all input-output operations are executed via a terminal. The package is also conversational in nature as it directs the user to the various stops of decision making by asking immediately relevant questions and responding intelligently to the user's answers. The user's response is always verified for accuracy. If the user responds incorrectly, a suitable feedback is immediately sent. All the necessary computations are done by the program itself and the user only need to concern with the logic of the decision making analysis.

**A Microcomputer Based Interactive System for User Defined Classification Algorithm Modelling**

Shvalb, V.

The considered system is a tool for defining, modelling and evaluating a user-defined classification (image classification) algorithm. Although the user of the system needs no experience in programming, an in-depth knowledge of the system application is necessary. The duration of training required for a system user is about a day. An example of application of the system includes choosing a set of significant factors in decision making.

The set of objects to be considered is described by a relational data base (kind of objects naturally depends on application). The attributes of the objects are referred as the components of image vectors.

A classification method is characterized by the choice of parameters, algorithms dealing with recoding of attributes, space metrics, and image transformation. Elementary descriptive statistics (averages, standard deviations, histograms, etc.) are used to improve the classification algorithm.

The system is partially menu-driven and partially command-driven. The image transformation and source data coding are handled by user's own written codes which can consist of command such as "if ...then ... else" and arithmetic and logical operations. The system is implemented on portable PC EPSON HX-20 and written in BASIC.

**A Decision Support System for Multiple Criteria Optimization**

Steuer, R. E., Silverman, J. & Whisman, A. W.

This paper discusses a decision support system (DSS) for solving multiple criteria optimization problems. Depending upon the problem, the DSS applies either the interactive weighted-sum or weighted Tchebycheff procedures to sequentially sample the set of nondominated criterion vectors. The DSS allows the user to modify the model while converging to a final solution. Computer graphics are used at the computer/user interface. Applications to multiple criteria optimization problems are discussed.