Strong Planning Revisited

Abstract

Strong planning under all state-action pairs has been researched in recent years. Normally, the more the state-action pairs are, the higher of complexity is. In fact, there are many state action pairs which are useless for strong planning. Before finding strong planning, it is significant to find a set of state-action pairs which are useless for strong planning, and to best of our knowledge, it is still an open problem. In this paper, for a nondeterministic state-transition system, describes a method about how to use the reach ability matrix to count the state reach ability in a nondeterministic state-transition system. Using the state reach ability, some important conclusions about strong planning are acquired. These conclusions tell us what state-action pairs are useless when we search strong planning. So a lot of state-action pairs can eliminate directly from all state-action pairs before searching strong planning. Our contribution is the some important conclusions about strong planning, these works are significant to improve the efficiency of algorithm for solving strong planning.