Analogical Replay: Utility Analyses
Case-based planning involves storing individual instances of problem-solving episodes and using them to tackle new planning problems. This paper is concerned with derivation replay, which is the main component of a form of case-based planning called derivational analogy (DA). Prior to this study, implementations of derivation replay have been based within state-space planning. We are motivated by the acknowledged superiority of partial-order (PO) planners in plan generation. Here demonstrate that plan-space planning also has an advantage in replay. We will argue that the decoupling of planning (derivation) order and the execution order plan steps, provided by partial-order planners, enables them to exploit the guidance of previous cases in a more efficient and straightforward fashion. We validate our hypothesis through a focused empirical comparison.