Preferential Semantics for Goals Planning to Achieve Goals Preferences and Utility

Goals, as typically conceived in AI planning, provide an insuucient basis for choice of action, and hence are deecient as the sole expression of an agent's objectives. Decision-theoretic utilities ooer a more adequate basis , yet lack many of the computational advantages of goals. We provide a preferential semantics for goals that grounds them in decision theory and preserves the validity of some, but not all, common goal operations performed in planning. This semantic account provides a criterion for verifying the design of goal-based planning strategies, thus providing a new framework for knowledge-level analysis of planning systems. In the predominant AI planning paradigm, planners construct plans designed to produce states satisfying particular conditions called goals. Each goal represents a partition of possible states of the world into those satisfying and those not satisfying the goal. Though planners use goals to guide their reasoning, the crude binary distinctions deened by goals provide no basis for choosing among alternative plans that ensure achievement of goals, and no guidance whatever when no such plans can be found. These lacunae pose signiicant problems for planning in all realistic situations, where actions have uncertain eeects or objectives can be partially satissed. To overcome these widely-recognized expressive limitations of goals, many AI planners make ad hoc use of heuristic evaluation functions. These augment the guidance provided by goals, but lack the semantic jus-tiication needed to evaluate their true eecacy. We believe that heuristic evaluation functions should not be viewed as mere second-order reenements on the primary goal-based representation of objectives, supporting a separate \optimizing" phase of planning. Our thesis is that relative preference over the possible results of a plan constitutes the fundamental concept underlying the objectives of planning, with goals serving as a computationally useful heuristic approximation to these preferences (Doyle, 1990). Our purpose here is to provide a formal semantics for goals in terms of decision-theoretic preferences that supports rational justiications for planning principles. The grounding in decision theory enables designers to determine whether their planning systems act rationally in accord with their goals, and provides a principled basis for integrating goals with other types of preference information. We begin by summarizing some basic concepts of preference. We then develop formal decision-theoretic semantics for goals and examine some standard planning operations in light of the semantics. We conclude by discussing some related work and ooering some directions for future investigation. Decision theory starts with …