mCP nets: representing and reasoning with preferences of multiple agents

Abstract

We introduce\nmCP nets, an extension of the CP net formalism\nto model and handle the qualitative and conditional preferences\nof multiple agents. We give a number of different\nsemantics for reasoning with\nmCP nets. The semantics are\nall based on the idea of individual agents voting. We describe\nhow to test optimality and preference ordering within a\nmCP\nnet, and we give complexity results for such tasks. We also\ndiscuss whether the voting schemes fairly combine together\nthe preferences of the individual agents.