This paper deals with preference representation and decision-making problems in the context of multiattribute utility theory. We focus on the generalized additive decomposable utility model (GAI) which allows interactions between attributes while preserving some decomposability. We present procedures to deal with the problem of optimization (choice) and ranking of multiattribute items. We also address multiperson decision problems and compromise search using weighted Tchebycheff distances. These procedures are all based on GAI networks, a graphical model used to represent GAI utilities. Results of numerical experiments highlight the practical efficiency of our procedures.