A fuzzy stochastic technique for project selection

The comparison of two or more risky projects is more of a challenge than the evaluation of one project in isolation. In the numerous decision models and methods suggested in the literature, often it is assumed that the criteria as well as the decision maker's preference or utility function can be crisply defined. Multi-attribute decision aids that permit the consideration of both multi-variables and risks generally have been associated with complex mathematics and heavy consumption of resources. This paper shows how project selection problems can be dealt with when some project attributes are subject to random variations. The method incorporates fuzzy analysis into multi-attribute utility theory. The aggregate utility function for an individual project is derived as a fuzzy number (or interval) which, in turn, yields probabilistic information for stochastic dominance tests. A unique feature of the approach is that it dispenses with the task of selecting probability distributions for aggregate utility functions. A comparison of the proposed method with the expected utility approach was made and the findings showed agreement between the results.