Model World: The Great Debate - MAUT Versus AHP

“The theory of decision analysis (DA) is designed to help the individual make a choice among a set of prespecified alternatives” (Keeney and Raiffa 1976, vii). Simply put, the basic DA problem is, How should a decision maker select from a given set of competing alternatives that are evaluated against conflicting objectives? Because operations research (OR) is the science of decision making, OR researchers and practitioners have developed a wide range of methods for resolving the basic DA problem and its variations. They base some of these methods on von Neumann and Morgenstern’s (1947) expected utility theory (EUT) and Savage’s (1954) subjective expected utility theory (SEUT), with extensions to multiattribute utility theory (MAUT) and multiattribute value theory (MAVT) (Keeney and Raiffa 1976, Edwards 1977, Edwards and Barron 1994). Other procedures they have developed and applied include Roy’s (1968, 1996) ELECTRE (election et choix traduisant la réalité), PROMETHEE (preference ranking organisation method for enrichment evaluations) by Brans et al. (1996), TOPSIS (technique for order preferences by similarity to ideal solutions) by Hwang and Yoon (1981), and Hwang et al. 1993), and the analytic hierarchy process (AHP) (Saaty 1977, 1980, 1982). People often refer to ELECTRE and PROMETHEE as representing the European School, and TOPSIS and the AHP (and others) as representing the American School. (Mollaghasemi and Pet-Edwards 1997 and Belton and Stewart 2002 describe most of these multicriteria decision-making (MCDM) methods.) Since its development 25 years ago, the AHP has become a widely popular MCDM procedure in the US and in other countries. The AHP has gained wide acceptance among academics and practitioners (Golden et al. 1989, Forman and Gass 2001, Golden and Wasil 2003). It is the workhorse for solving multicriteria problems that have a finite number of alternatives to be put in rank order by priority (weight). The AHP seems to have replaced MAUT and MAVT for solving such real-world MCDM problems. As an exponent of the AHP, I introduced it to my MBA and undergraduate management science students in the early 1980s and was the first to discuss it in a text (Gass 1985). I am an AHP fan. But as the new kid on the block, the AHP has had to compete against established methods, in particular, MAUT. Many academics schooled in MAUT debated the efficacy of the AHP. In the mid-1980s, MCDM researchers began to compare the AHP to MAUT. They found that the AHP does not adhere to the basic von Neumann and Morgenstern axiomatic structure of normative utility theory as incorporated in MAUT, and they raised other concerns. I tried to reconcile the criticisms of the AHP with my understanding of AHP and MAUT and joined E. Forman in describing our view of the debate (Forman and Gass 2001). We concluded: