Multi-Objective Variable Elimination for Collaborative Graphical Games

In this paper we propose multi-objective variable elimination (MOVE), an efficient solution method for multi-objective collaborative graphical games (MO-CoGGs), that exploits loose couplings. MOVE computes the convex coverage set, which can be much smaller than the Pareto front. In an empirical study, we show that MOVE can tackle multi-objective problems much faster than methods that do not exploit loose couplings.