RIA: Developing Reliable Plan Knowledge Bases

虽然人工智能规划的研究人员 我已经关注故障恢复和调试一段时间了， 还没有一致的努力来评估故障恢复， 规划或支持规划知识的设计和调试 新环境的基础。 拟议的研究填补了这一空白 分析规划系统的新工具和 确定这些系统失败的原因以及可以做些什么。 PI的论文描述了一个过程，故障恢复分析 (FRA)，用于分析规划者（凤凰 规划师）在其环境中，以确定是否以及如何 导致自己的失败。 私家侦探建议将她的论文 通过将其应用于另一个规划者和环境进行研究， 在工具中自动化和集成FRA，以帮助设计师 评估和调试规划器知识库。//

While researchers in artificial intelligence planning have been concerned with failure recovery and debugging for some time, there has been no concerted effort to evaluate failure recovery and planning or support the design and debugging of plan knowledge bases for new environments. The proposed research fills this gap with new tools for analyzing planning systems and a methodology for determining why these systems fail and what can be done about it. The PI's thesis described a procedure, Failure Recovery Analysis (FRA), for analyzing the performance of a planner (the Phoenix planner) in its environment to determine whether and how it might be causing its own failure. The PI proposes to extend her thesis research by applying it to another planner and environment, and automating and integrating FRA in a tool to assist designers in evaluating and debugging planner knowledge bases.//