Improving preference elicitation for decision support systems

Preferences form the input and control the output of all complex decision making processes. Whether they appear in the form of objective functions, goals, or criteria, preferences and their accurate elicitation are crucial to the formulation of sound decision models both in theory and in practice. Unfortunately, common methods for eliciting preferences are flawed. In this paper, alternative preference elicitation methods for multiple criteria decision making (MCDM) are reviewed and an improved iterative, interactive, two-method approach is proposed for a decision support system design. It is believed that using the proposed design takes advantage of all the merits of each of the two components, and helps the decision maker explore the elements of a problem by bringing inconsistencies in the preference structure to the forefront for resolution