Declarative and procedural goals in intelligent agent systems

摘要: An important concept for intelligent agent systems is goals. Goals have two aspects: declarative (a description of the state sought), and procedural set plans achieving goal). A view goals necessary in order to reason about properties goals, while a ensure that can be achieved efficiently dynamic environments. In this paper we propose framework which integrates both views. We discuss requisite link between aspects, then derive formal semantics has these properties. present high-level plan notation with give its semantics. show how use information permits reasoning (such as detection resolution conflicts) performed on