Incorporating Fuzzy Components Into Metric Graphplan

People's researching on planner has lasted for nearly fifty years. In all planners, the most famous one was the classical Graphplan which was developed by Avirm Blum and Merrick Furst in 1995. Graphplan was the first planner which using the high efficiency planning graph structure in planner design. From then on, many good planners inherited this structure. But Graphplan has not break through the hard constrains of STRIP problem expression and three supposes of classical planning. So it could not express and solve planning problems which contains resources distribution and subtle information fetching.Planning which distributed resource to planning objects by arithmetric expression was called metric planning. Extra resource constrains made it was so difficult that resolve metric planning problems on graph structure. In 1999, Jana Koehler developed metric Graphplan which extended the problem expression of Graphplan and could solve both classical and metric problems. But metric Graphplan inherits Graphplan's high efficiency as well as its hard constrained framework. It is argued herein that this framework is too rigid to capture the subtlety of many real problems.Most planners fetched subtle information by soft constrains which included preference constrains and priority constrains. They were both defined on the planning objects which related with the planning problem and its measurement reflected their importance in planning. Fuzzy sets system which was defined by Dubios and could expression the uncertain problems supported nature measure tools for preference constrains and priority constrains.A new metric Graphplan algorithm that incorporates fuzzy components into metric Graphplan and extracts plan by incremental local change is defined. Comparatively, fuzzy components extend the preference and priority expression of metric Graphplan, and softened hard constrains to soft constrains. A new plan algorithm based on incremental local change is used to control nodes booming and improving search efficiency.