Abstract: On purpose to improve metric Graphplan’s capability that captures the detail information in the real world, a fuzzy component model was constructed which was defined by preference and priority merbership functions, and fuzzy components were incorporated into metric graphplan by adding preference and priority fields in problem expression. For viewing the process of searching for a plan as a constraint satisfaction problem, we used incremental local change algorithm to improve the efficiency of problem resolving. The algorithmmostly reused the previous problem resolving results that could avoid much repeat, and support a theoretical foundation to develop efficiency metric planner.

Key words: metric graphplan, graphplan, preference, priority, fuzzy component