Breakthroughs in Decision Science and Risk Analysis

Chapter 5Using Preference Orderings to Make Quantitative Trade-Offs

Chen Wang¹ and Vicki M. Bier²

¹ Department of Industrial Engineering, Tsinghua University, Beijing, P.R., China

² Department of Industrial and Systems Engineering, University of Wisconsin-Madison, Madison, WI, USA

Introduction

Real-world decision-makers often face multiple objectives. For example, many decisions in the business context aim to maximize both profits and market share (Keeney, 2007), while consumers may want to trade off the quality of a product against its price. Multiattribute utility theory (MAUT; Keeney and Raiffa, 1976) is one of the most well-known and rigorous methods for trading off multiple (generally conflicting) objectives. The additive form of the multiattribute utility function is often used, due to its simplicity, and has been shown to give satisfactory results in a wide variety of circumstances (Dawes and Corrigan, 1974; Stewart, 1996).

One important task in constructing multiattribute utility functions is to choose appropriate attributes to quantitatively reflect stakeholder objectives. Desirable attributes should be measurable on either natural scales (e.g., lives saved, dollars spent) or constructed scales. For an example of a constructed scale, Keeney and von Winterfeldt (1994) use a 0–5 scale to measure the environmental impact of nuclear waste disposal strategies. Attributes should also directly and unambiguously describe the degree to which the decision-maker’s objectives are met (Clemen ...