GAMER: Fully-Observable Non-Deterministic Planning via PDDL-Translation into a Game

This paper presents an optimal planner for the international probabilistic planning competition at ICAPS-08, IPPC-2008 for short. The planner solves non-deterministic action planning problems with binary decision diagrams. The efficiency of the planning approach is based on a translation of the non-deterministic planning problems into a twoplayer turn-taking game, with a set of actions selected by the solver and a set of actions taken by the environment. The formalism we use is a PDDL-like planning domain description language that has been derived to parse and instantiate general games. This conversion allows to derive a concise description of planning domains with a minimized state vector, thereby exploiting existing static analysis tools for deterministic planning. Subsequently, we apply strong and strong cyclic planning algorithms as found in the literature. We additionally observe that the policy can be extracted naturally.