Compiling Causal Theories to Successor State Axioms and STRIPS-Like Systems

摘要

We describe a system for specifying the effects of actions. Unlike thosecommonly used in AI planning, our system uses an action description languagethat allows one to specify the effects of actions using domain rules, which arestate constraints that can entail new action effects from old ones.Declaratively, an action domain in our language corresponds to a nonmonotoniccausal theory in the situation calculus. Procedurally, such an action domain iscompiled into a set of logical theories, one for each action in the domain,from which fully instantiated successor state-like axioms and STRIPS-likesystems are then generated. We expect the system to be a useful tool forknowledge engineers writing action specifications for classical AI planningsystems, GOLOG systems, and other systems where formal specifications ofactions are needed.