Constructing Belief Networks to Evaluate Plans
This addresses plan evaluation in AI planning systems, but appears incremental as it builds on existing belief network methods for specific plan features.
The paper tackles the problem of constructing belief networks to evaluate plans from a knowledge-based planner, presenting techniques to handle complicating features like context-dependent consequences, contingencies, and multiple agents, but does not report concrete numerical results.
This paper examines the problem of constructing belief networks to evaluate plans produced by an knowledge-based planner. Techniques are presented for handling various types of complicating plan features. These include plans with context-dependent consequences, indirect consequences, actions with preconditions that must be true during the execution of an action, contingencies, multiple levels of abstraction multiple execution agents with partially-ordered and temporally overlapping actions, and plans which reference specific times and time durations.