Skip to main content

Über dieses Buch

The pioneering study by Bowman [1980) reawakened interest in risk and return relations in the strategic management literature. We do not examine this literature here because we have elsewhere reviewed it in detail 1 and because, for the most part, these studies have been confined to ex post data. Discussions of the strategies which subjects used to direct their ex ante evaluations of risks and returns have either been omitted or else have been only indirectly inferred from ex post data. In addition, with few exceptions, this literature does not attempt to ascertain the meanings that might have been assigned by subjects to terms like "risk" and/or the "returns" with which they have been concerned. Even fewer of these studies have attempted to ascertain how the subjects implemented their definitions en of prospective strategies. Thus, tius literature may route to arriving at evaluations best be regarded as bearing only indirect relations to the present study which is concerned not only with the meanings assigned to terms like "risk" and "return" but also with how these terms are used in arriving at risk and return evaluations of proposed strategies as well as how they are measured and used, on an ex ante basis en route to seeing how these evaluations match with ex post performance. In a sense, one part of this study--i. e.




Quot homines, tot sententiae

This volume is a Festschrift in honour of Stanley Zionts. The occasion is his upcoming 60th birthday in January, 1997. Dr. Zionts, Alumni Professor of Decision Support Systems, State University of New York at Buffalo, is one of the leaders of the Field of Multiple Criteria Decision Making. He is the founder of the International Society of Multiple Criteria Decision Making and its first President. In the seventeen years since its formation, under his leadership and under the leadership of Professor Ralph Steuer, the Society has grown and become very international, reflecting the nature of the field. Today it encompasses some 1200 members representing 82 different countries.
Mark H. Karwan, Jaap Spronk, Jyrki Wallenius

Multiple Objective Linear Programming and Interactive Methods


Convergence and Validation of Interactive Methods in MCDM: Simulation Studies

A wide array of interactive methods, or (more correctly perhaps) methods of progressive articulation of preference, are available to the MCDM practitioner. In some cases (such as the Zionts-Wallenius method) it can be proved that the procedures converge in the sense of terminating after a finite number of iterations. In other instances, no such convergence can be demonstrated, and such methodologies must be classed as heuristics. It is, however, by no means evident that mathematical convergence is either necessary or sufficient to demonstrate the practical validity of a procedure. Practical validity requires that within a very small number of iterations, the decision maker is enabled to gain sufficient understanding of the decision Space and the necessary trade-offs, to have confidence in and be satisfied with the Solution selected at that stage with the help of the method. In this paper we review Simulation studies, in which interactive methods of both the value function and goal programming types have been implemented in hypothetical computer-generated decision contexts. The results of these Simulation studies provide a substantial level of Validation for both types of interactive methods, but do also provide warnings of how careless implementation of these methods can lead to very poor results.
Theodor J. Stewart

Convergence of Interactive Procedures of Multiobjective Optimization and Decision Support

The paper presents an overview of issues of convergence of inter-active procedures in multiobjective optimization and decision support. The issue of convergence itself depends on assumptions concerning the behavior of the decision maker — who, more specifically, is understood as a user of a decision support system. Known procedures with guaranteed convergence under classic assumptions are reviewed. Some effective procedures of accelerated practical convergence but without precise convergence proofs are recalled. An alternative approach to convergence based on an indifference threshold for increases of value functions or on outranking relations is proposed and illustrated by a new procedure called Outranking Trials.
Andrzej P. Wierzbicki

A Hybrid Interactive Technique For The MCDM Problems

An interactive technique proposed in this paper allows the decision maker to use different search principles depending on his/her perception of the achieved values of the objectives and trade-offs. While an analysis of values of the objectives may guide the initial search for a satisfactory Solution, it can be replaced by trade-off evaluations at some later stages of interactive decision making.
Ignacy Kaliszewski, Wojtek Michalowski, Gregory Kersten

A Comparison of Aspiration Level Interactive Method (AIM) and Conjoint Analysis in Multiple Criteria Decision Making

The predictive validity of the two methods, Aspiration-level interactive method (AIM) and conjoint analysis, used for solving decision problems involving discrete alternatives are compared. An empirical analysis based on subjects’ preferences for a multiattribute product (buying a house) and a service (selecting an MBA program for study) indicated that consumer preferences derived from AIM may be more valid than the preferences derived from the full-profile conjoint analysis method.
Madhukar Angur, Vahid Lotfi

Reference Direction Approach to Multiple Objective Linear Programming: Historical Overview

The purpose of this paper is to provide a historical overview and state-of-art review of the development of the reference direction approach to multiple objective linear programming, originally proposed by Korhonen and Laakso [1986a], We will discuss the original approach and several development phases. First, the basic idea was extended by allowing a dynamic specification of the reference direction. This extension, which enables a decision-maker to freely search nondominated solutions, is called Pareto Race by Korhonen and Wallenius [1988], Pareto Race is an essential part of the decision support system VIG, developed by Korhonen [1987a], to support the solving of multiple objective linear programming problems. Furthermore, several other development phases are reviewed.
Pekka Korhonen

Implementing the Tchebycheff Method in a Spreadsheet

The Tchebycheff Method [29] is one of a group of interactive procedures [e.g., 3–7, 10, 13–14, 16–19, 21, 23, 25-27, 30–31] for solving multiple objective programming problems. While the procedures in this group represent a diversity of Solution strategics, the implementations of most of these procedures posses remarkable similarities [8, 9], with most of these procedures being essentially supervisory routines designed to call commercial-grade linear, integer and nonlinear programming codes as their workhorse Software. That is, to probe the Solution set at each iteration, one or more (single objective) mathematical programs must be solved, at which point in earlier Computing environments, an appropriate solver such as MINOS (Murtagh and Saunders [1977]) for linear problems, MPSX (IBM [1979]) for integer problems, or GRG2 (Lasdon and Waren [1986]) for nonlinear problems would be called.
Ralph E. Steuer

Preferences and Learning


From Maximization to Optimization: MCDM and the Eight Models of Optimality

Since its inception, Multiple Criteria Decision Making (MCDM) was destined to evolve alternative notions of optimality. Its challenges to the traditional OR/MS single-objective maximization have crossed the generations, continuing for some 25 years. Yet, its alternative notions of optimality still remain mostly undeveloped.
It is multiple rather than single criteria that characterize “the best” (optimal) in the areas of economics, engineering, management and business. In this paper we develope the notion of optimum conceived as a balance among multiple criteria. We propose a classificational scheme of eight different, separate and mutually irreducible optimality concepts, with the traditional single-objective “optimality” representing a one special case. The eight optimality concepts provide a useful foundation and framework for the future of MCDM in both research and applications.
Milan Zeleny

A General Model of Preference Aggregation

A general model is presented which encompasses many procedures used for aggregating preferences in multicriteria decision making (or decision aid) methods. Are covered in particular: MAUT, ELECTRE and several other outranking methods. The main interest of the model is to provide a key for understanding the differences between methods. Methods are analyzed in terms of their way of dealing with “preference differences” on each criterion/attribute. The more or less large number of equivalence classes of preference differences that can be distinguished in a method helps to situate it in a continuum going from compensatory to noncompensatory procedures, from cardinal to ordinal methods.
D. Bouyssou, M. Pirlot, Ph. Vincke

Issues In Supporting Intertemporal Choice

Intertemporal decision making refers to contexts in which the consequences accumulate in stages over time. Attention is confined to cases in which the stages are discrete. The purpose of this paper is to highlight some of the anomalies between how people‘ should’ make and how they ‘do’ make intertemporal decisions: i.e. between findings in the normative and descriptive literatures. The paper indicates some of the framing issues which must be considered when trying to obtain a decision maker’s preferences, especially with respect to the way the questions are posed. The intention is to identify some of the bridges which need to be built between descriptive and normative ideas if decision makers are to be supported effectively in making intertemporal decisions.
The concepts and ideas are illustrated in the context of the decisions faced after a nuclear accident to indicate how information should be presented in order to obtain ‘unbiased’ preferences from decision maker - or, at least, to reduce any bias. These issue are currently being faced in the design of RODOS, a decision support system being built to aid emergency management in the event of a nuclear accident with off-site and, hence, probable long term consequences.
Elizabeth Atherton, Simon French

Cancellation Conditions for Multiattribute Preferences on Finite Sets

Applications of decision theory for multiple criteria or multiple attributes often assume a Utility representation for preferences that is additive over criteria or attributes. Axiomatic theories for additive Utilities are well developed but are not without gaps. A case in point arises with finite sets of alternatives, where two preference axioms are necessary and sufficient for additive Utilities. One is weak ordering. The other is a cancellation ax-iom that consists of an infinite scheme of cancellation conditions, one for each positive integer K ≥ 2. It is known that the infinite scheme can be truncated to a finite scheme for KK* that depends on the size of the set of alternatives, but very little is known about the value of K* which ensures additivity for all finite sets of that size. The present paper contributes to the determination of K*. A fundamental result is that if there are m attributes, the j th of which has n j elements in its attribute set, then Σn j - (m - 1) is an upper bound on K*. Lower bounds on K* that are near to this upper bound are obtained for special cases of (n1, n2, ..., nm.
Peter C. Fishburn

Measurement of Differences in Preferences Expressed by a Simple Additive Rule

The Method of Weighted Sinn of Criteria Estimates (WSCE) is popular in multicriteria decision making. The question arises, how to measure differences in the preferences of different decision makers or the same decision maker (DM) over time when using the WSCE method. A new measure is suggested:the number of alternative pairs in complete order given by WSCE for which the superiority of one alternative upon the other depends 011 DM’s preferences.They are called as potentially contradictory pairs (PCA). It is shown that PCA created from the elements of the first Pareto-layer very reliably represent PCA in a complete order in the case of additivity.This makes it possible to only use PCA to measure differences in DM’s preferences.
O. I. Larichev

A Foundation of Principles for Expanding Habitual Domains

A human brain is like a super Computer (Hardware) and has practically unlimited potential capacity. However, because of their habitual patterns of thinking (Software), people use only a very small percentage of the capacity. Their ability for effective problem-solving very much relies on how they Up-grade the Software and break undesirable and inefficient habits. Habitual domain studies aim at providing general principles, techniques, and theoretical foundations of how to expand and enrich habitual domains. In this paper, we show the importance of studying habitual domains using familiar examples. We introduce basic concepts of the habilual domain theory. We present three of the most commonly used principles of expanding habitual domains. Finally. as a new conttibution of this paper. we present a theoretical Interpre-tation of the principles using the notion of activation probabilities and attention spectra.
Po-Lung Yu, Liping Liu

Exploration of Multi-objective Problems in Artificial Intelligence

One of main themes in artificial intelligence is to simulate brains of human beings. This can not be almost realized by only a model of brain, because brains of human beings have a great number of functions and perform them in complex manners, Therefore, techniques in artificial intelligence are usually discussed for those functions separately- For example, decision making is a major function of brain. Since it seems not to be practical to automatize decision making by artifical intelligence, this function has been studied from a viewpoint of decision support by a system based on collaboration of machines (Computers) and human beings. Interactive multi-objective programming techniques are good examples in this field.
Hirotaka Nakayama



Design Decomposition in Quality Function Deployment

Quality function deployment (QFD) is a cross-functional planning tool which ensures that the voice of the customer is systematically deployed throughout the product planning and design stages. One of the impediments to the adoption of QFD as a product design aid has been the large size of the design matrix, called a house of quality (HOQ) chart. This paper presents a formal approach to reducing the size of a HOQ chart using the concept of design decomposition combined with multiattribute value theory. The proposed approach aims to decompose a HOQ chart into smaller subproblems which can be solved efficiently and independently. The smaller size of the subproblems would make the QFD process simpler, easier, and cheaper. The decomposition algorithm proposed in this paper can also be applied to forming machine cells and part families in flexible manufacturing systems when the design incidence matrix is nonbinary with no diagonal structure and the column entities are correlated.
Kwang-Jae Kim, Herbert Moskowitz, Jong-Seok Shin

A Study Of Evaluations Of Mutual Fund Investment Strategies

Evaluations of publicly announced investment strategies of open-end mutual funds were secured from a questionnaire sent to brokers and editors of financial newsletters. Ex Ante evaluations secured in this manner are compared with ex post performance over the period 1984-1988 and rated via data from Morningstar, Inc., and Business Week on mutual fund Performances over this same period. The present study is designed so that separations can be effected between risk and return evaluations and Performance over all pertinent ex ante and ex post pairings. Ex ante evaluations of risk and return are found to be positively correlated—as posited in the finance, decision theory and economics literatures—but their ex post pairings are negatively related. Somewhat surprisingly, ex ante to ex post evaluations of risk are positively correlated while ex ante to ex post evaluations of return are negatively correlated. The source of the ex ante to ex post negative relations is therefore in the return rather than the risk evaluations. This casts additional light on the Bowman Paradox (and related topics) which have been extensively studied in the Strategic management literature. It also adds to these paradoxes and suggests programs for further research in a manner that is discussed in the addendum to this paper.
Patrick L. Brockell, W. W. Cooper, Ku-Hyuk Kwon, Timothy W. Ruefli

A Multi-Dimensional Framework for Portfolio Management

Given an opportunity set of financial securities, an investor faces the problem of selecting one portfolio out of the multiplicity of portfolios that can be formed. This paper presents a multi-dimensiona! framework that can serve as a decision aid in the portfolio selection and management process. The framework yields room for different settings of the portfolio management problem and offers an alternative to both unstructured ad hoc approaches and complex approaches that severely restrict the decision process. The framework offers the investor systematic guidance in the search for a portfolio that suits his tastes and preferences best, while satisfying the restrictions he faces.
Winfried Hallerbach, Jaap Spronk

MCDM and Sustainable Development: The Case of Water Resources Planning in India

Sustainable development has emerged as the centerpiece of natural resources management and environmental protection. Defined as the use of natural resources today so as not to compromise the ability of future generations to use them, sustainable development is a compelling idea, and of indisputable importance. Translating it into plans and actions, however, is very difficult both because of the value conflicts inherent in sustainable development and the need for new planning and analytical methodologies.
Jared L. Cohon

Role of BATNA and Pareto Optimality in Dyadic Multiple Issue Negotiations

The role of the Best Alternative to a Negotiated Agreement (BATNA) has been the subject of much discussion in the literature recently. We extend this discussion concerning BATNAs and reservation prices to multiple issues. We have conducted an experiment to test several hypotheses concerning the role of BATNAs in a contextually relevant three-issue dyadic negotiation. We further improve the methodology of conducting negotiation experiments by not requesting the subjects to follow a given value point structure. We report the results of the experiment and discuss their implications.
Jeffrey Teich, Pekka Korhonen, Jim Phillips, Jyrki Wallenius

Multiple Criteria Scheduling on a Single Machine: A Review and a General Approach

Scheduling problems have been treated as single criterion problems until recently. Many of these problems are computationally hard to solve even as single criterion problems. However, there is a need to consider multiple criteria in a real life scheduling problem in general. In recent years, many articles considering bicriteria scheduling problems, and some articles considering more than two criteria started to appear in the literature. We briefly review various categories of approaches on multiple criteria scheduling on a single machine. We then present a general approach to find the most preferred schedule in a bicriteria environment for any given nondecreasing composite function of the criteria.
Murat Köksalan, Suna Köksalan Kondakci

Multiobjective Covering and Routing Problems

In this paper, we review research on point location covering problems and their development from the Maximal Covering Location Problem. With this as a basis, a particular path problem (the Maximal Covering Shortest Path Problem) is chosen to serve as a prototype combined covering - location - routing problem and possible developments are suggested. Finally, we discuss possible avenues for further research.
Brian Boffey, Subhash C. Narula

An Evaluation of Our National Policy to Manage Nuclear Waste from Power Plants

The current national policy to manage nuclear waste from power plants is to dispose of it in a repository to be constructed at Yucca Mountain in Nevada. With many different assumptions about uncertainties and objectives, this strategy is shown to be the equivalent of $10,000 million to $50,000 million inferior to other available strategies. The implications of the analysis strongly suggest that our national policy to manage nuclear waste should be changed.
Ralph L. Keeney, Detlof von Winterfeldt

Professor Stanley Zionts

Fifteen Years of Work with a Guru: A Lifetime or Just the Beginning

The sixtieth birthday celebration of Stan Zionts is an important event in the lives of many - his numerous students and colleagues who have benefited enormously through his association. I have had the unique privilege of a sustained association with him for the past one and a half decades, and being the beneficiary of his mentorship, comradeship and collaboration in numerous research ventures. In this article, I summarize in a story-like fashion, our research association, accomplishments and the road ahead.
R. Ramesh
Weitere Informationen