Formalizing Plan Justi cations

This paper formalizes the notion of justiied plans, which captures the intuition behind \good" plans. A justiied plan is one that does not contain operators which are not necessary for achieving a goal. The importance of formalizing this notion is due to two reasons. First, it gives rise to methods for optimizing a given plan by removing \useless" operators. Second, several important concepts describing abstraction hierarchies are deened via justiied plans. In the past, relatively few attempts have been made to formalize such a notion. This paper deenes several diierent kinds of plan justiica-tions, presents algorithms for nding a justi-ed version of a plan, and shows that the task of nding the best possible justiied version of a plan is NP-complete. Finally, it presents a greedy algorithm for nding a near-optimal jus-tiied plan in polynomial time.