Remarks on the Analytic Hierarchy Process

Abstract

The analytic hierarchy process (AHP) is flawed as a procedure for ranking alternatives in that the rankings produced by this procedure are arbitrary. This paper provides a brief review of several areas of operational difficulty with the AHP, and then focuses on the arbitrary rankings that occur when the principle of hierarchic composition is assumed. This principle requires that the weights on the higher levels of a hierarchy can be determined independently of the weights on the lower levels. Virtually all of the published examples of the use of the AHP to evaluate alternatives relative to a set of criteria have assumed this principle. The key to correcting this flaw is the synthesis of the AHP with the concepts of multiattribute utility theory.decision analysis, multiattribute utility theory, analytic hierarchy process

Does AHP help us make a choice? An experimental evaluation

“…Several papers have compiled the numerous AHP success stories (Forman and Gass 2001;Golden et al 1989;Ho 2008;Kumar and Vaidya 2006;Liberatore and Nydick 2008;Omkarprasad and Sushil 2006;Saaty and Forman 1992;Shim 1989;Vargas 1990;Zahedi 1986), but its popularity does not verify that the AHP recommendation is always the best alternative. In fact, AHP has been sharply criticised on several points (Bana e Costa and Vansnick 2008;Barzilai 2001;Belton and Gear 1983;Dodd and Donegan 1995;Donegan et al 1992;Dyer 1990;Holder 1991;Johnson et al 1979;Pöyhönen et al 1997;Salo and Hamalainen 1997;Webber et al 1996). Many papers have theoretically compared or at least grouped multi-criteria decision methods by similarities (Al-Shemmeri et al, 1997;Guitouni and Martel, 1998;Guitouni et al, 2007;Kornyshova and Salinesi, 2008;Simpson, 1996).…”

A fuzzy group Electre method for electronic supply chain management framework selection

“…Harker and Vargas (1987) show that the AHP does have an axiomatic foundation, the cardinal measurement of preferences is fully represented by the eigenvector method, and the principles of hierarchical decomposition and rank reversal are valid. Conversely, Dyer (1990a) has questioned the theoretical basis underlying the AHP and argues that it can lead to preference reversals rather than the best or preferential outcome. In response, Saaty (1990b) explains how rank reversal is a positive feature when new reference points are introduced.…”

“…In response, Saaty (1990b) explains how rank reversal is a positive feature when new reference points are introduced. In this study, the geometric aggregation rule is used to avoid the controversies associated with rank reversal (Dyer 1990a, Harker and Vargas 1990, Saaty 1990b). …”

“…The process, as described in Section 2.1, is simplified by confining the estimates to a series of pairwise comparisons. We aggregate the individual voting power weights into the group weights with the geometric aggregation rule to avoid the controversies associated with rank reversal (Dyer 1990a, Harker and Vargas 1990, Saaty 1990b.…”

Seleção dos objetivos fundamentais de uma rede de cooperação empresarial

“…Segundo Carvalho (1997), apesar do amplo e crescente espectro de aplicações do AHP, existem algumas restrições quanto ao uso deste método, que podem ser aprofundadas nos trabalhos de Dyer (1990a;1990b) e Saaty (1990). Um dos problemas apontados no AHP é a quantidade de comparações paritárias necessárias, que cresce muito rapidamente com o tamanho da matriz destas comparações.…”