Planning as Satisfiability with Relaxed $-Step Plans

Planning as satisfiability is a powerful approach to solving domain independent planning problems. In this paper, we consider a relaxed semantics for plans with parallel operator application based on -step semantics. Operators can be applied in parallel if there is at least one ordering in which they can be sequentially executed. Under certain conditions, we allow them to be executed simultaneously in a state s even if not all of them are applicable in s. In this case, we guarantee that they are enabled by other operators that are applied at the same time point. We formalize the semantics of parallel plans in this setting, and propose an effective translation for STRIPS problems into the propositional logic. We finally show that this relaxed semantics yields an approach to classical planning that is sometimes much more efficient than the existing SAT-based planners.

• Type: Conference
• Year: 2007
• Where: AUSAI
• Authors: Martin Wehrle, Jussi Rintanen
• Keywords: Artificial Intelligence, AUSAI 2007, Operators, Parallel Operator Application, Relaxed Semantics