Problem-Structure-Based Pruning for Cost-Optimal Planning

In this paper we focus on efficient methods for pruning the state space in costoptimal planning. The use of heuristics to guide search and prune irrelevant branches has been widely and successfully explored. However, heuristic computation at every node in the search space is expensive, so reducing the number of nodes to be considered by up-front analysis has great potential. Our contributions are not concerned with heuristic guidance, rather completeness-preserving pruning techniques that reduce the number of states a planner must explore to find an optimal solution. We focus on search space reduction and aim to draw the attention of the optimal planning community towards pruning techniques based on problem analysis that are orthogonal to the use of heuristics. We present results showing that our techniques can improve upon state-of-the-art optimal planners, both when using blind search and importantly in conjunction with modern heuristics, thus improving their potential.