Using temporal logics to express search control knowledge for planning

摘要: Over the years increasingly sophisticated planning algorithms have been developed. These made for more efficient planners, but unfortunately these planners still suffer from combinatorial complexity even in simple domains. Theoretical results demonstrate that is worst case intractable. Nevertheless, particular domains can often be tractable by utilizing additional domain structure. In fact, it has long acknowledged independent need dependent information to help them plan effectively. this work we present an approach representing and specific control knowledge. particular, show how search knowledge represented a temporal logic, then utilized effectively forward-chaining planner. There are number of advantages our approach, including declarative semantics knowledge; high degree modularity (new added without affecting previous knowledge); independence details algorithm. We implemented ideas TLPLAN system, able its remarkable effectiveness wide range