Yhteenveto: | This study focuses on the descriptive and prescriptive modelling of preferences within a discrete alternative multiple criteria decision framework. Pairwise preference comparisons of decision alternatives are a basic input of many multiple criteria decision models presented in the literature. This is the primary means of collecting preference information from the decision-maker (DM) in this dissertation. This study contains five chapters: the introduction and four research chapters. The four research chapters discuss the following problems: (1) Acquiring binary preference information for multiple criteria decision models (2) Generalizing prospect theory to the multiple criteria decision-making context (3) Solving the discrete multiple criteria problem using linear prospect theory (4) Further developments and tests of a progressive algorithm for multiple criteria decision making. The first two research chapters focus on descriptive multiple criteria choice behavior. The results indicate that existing value function-based decision models rely on preference information which the DM may not consistently be able to supply. Instead, support is found for the Kahneman-Tversky' s prospect theory as a reasonable model of choice for many individuals also in the multiple criteria decision context. In the third research chapter we develop an interactive method based on pairwise preference information for eliminating inferior alternatives with prospect theory-type value functions. This work is based on theory of convex cones developed by Korhonen, Wallenius, and Zionts. The fourth research chapter includes various tests, a computer implementation, and further developments of the progressive algorithm published by Korhonen, Moskowitz and Wallenius.
|