LIFT-UP: Lifted First-Order Planning Under Uncertainty

We present a new approach for solving first-order Markov decision processes combining first-order state abstraction and heuristic search. In contrast to existing systems, which start with propositionalizing the decision process and then perform state abstraction on its propositionalized version we apply state abstraction directly on the decision process avoiding propositionalization. Secondly, guided by an admissible heuristic, the search is restricted to those states that are reachable from the initial state. We demonstrate the usefulness of the above techniques for solving first-order Markov decision processes within a domain dependent system called FluCaP which participated in the probabilistic track of the 2004 International Planning Competition. Working toward a domain independent implementation we present novel approaches to θ-subsumption involving literal and object contexts.