Abstract
The interest in formal methods for handling multiple-criteria problems has increased during the last few years. A number of different approaches have been suggested but only a few implementations have been reported. In this dissertation several multiple-criteria methods based on utility theory are analyzed and compared. These methods are used to form a model of the decision-maker’s preferences, so that a utility index will be obtained for every alternative. Some of these methods establish the decision-maker’s preferences through a direct questioning (direct methods), while others derive them from his past choices or from his evaluations of a set of alternatives (indirect methods). In addition to studying the methods of estimation various models of the decision-maker’s preferences are analyzed. The methods and models are examined in a real-life problem. The attributes studied include accuracy, the decision-maker’s belief in the models, time requirement, ease of using the methods and models, etc. The experiment indicated that simple linear and additive models were found to be superior to the more complex models on these attributes. The precision of the models derived by the direct methods was also found to be lower than that of the models estimated by the indirect methods. The effect uncertainty in the estimates given by the decision-maker has on the selection of the multiple-criteria model was investigated in a simulation study. The results indicated that in very many situations where a more complex model than the additive is chosen, a model with greater precision would in fact be attained if the additive model was chosen instead.