Multiple objective evaluation and prioritization under risk with partial preference information†

Abstract

Much current interest surrounds the design and development of interactive multi-criterion decision support processes. We consider a finite alternative course of action, a single-stage, decision situation model under risk with multiple attributes, criteria or objectives. In this approach, a decision-maker is able to select interactively and iteratively a mixture of complete judgement (e.g. intuitive affect), heuristic elimination (e.g. elimination by aspects), and holistic assessment (e.g. expected utility theory) of partial preference information to enable aggregation of some, but not necessarily all, criteria or attributes. We consider three dominance relations on the alternative set : expected-value score dominance, first-order stochastic dominance, and second-order stochastic dominance. Two important issues are resolved here. The first of these involves establishment of the impact of various forms of partial objective aggregation information, in the form of attribute weights, upon dominance relations f...