Plan Understanding: Inferring Implicit Dependencies from Explicit Elements in Multi-Agent Plan Representations

Current planning systems often fail to represent the reasons why certain planning decisions are made. Explicit representation of this Plan Rationale is crucial for automated plan monitoring systems. For the past year we have been developing approaches to infer missing plan information from explicit plans. This paper presents a detailed view of our approach to inferring one type of Plan Rationale, Plan Interdependencies, which describes the supportive relationships between plan activities. We describe our application of this approach to real-world multiagent plan representations.