Learning from Goal Interactions in Planning: Goal Stack Analysis and Generalization

Natural Language

Volume

Proceedings of the AAAI Conference on Artificial Intelligence, 10

Planning

This paper presents a methodology which enables the derivation of goal ordering rules from the analysis of problem failures. We examine all the planning actions that lead to failures. If there are restrictions imposed by a problem state on taking possible actions, the restrictions manifest themselves in the form of a restricted set of possible bindings. Our method makes use of this observation to derive general control rules which are guaranteed to be correct. The overhead involved in learning is low because our method examines only the goal stacks retrieved from the leaf nodes of a failure search tree rather than the whole tree. Empirical tests show that the rules derived by our system PAL, after sufficient training, performs as well as or better than those derived by systems such as PRODIGY/EBL and STATIC.