GAI Networks for Decision Making under Certainty

This paper deals with preference elicitation and preference-based optimization in the context of multiattribute utility theory under certainty. We focus on the generalized additive decomposable utility model which allows interactions between attributes while preserving some decomposability. We first present a systematic elicitation procedure for such utility functions. This procedure relies on a graphical model called a GAI-network which is used to represent and manage independences between attributes, just as junction graphs model independences between random variables in Bayesian networks. Then, we propose an optimization procedure relying on this network to compute efficiently the solution of optimization problems over a product set.