accelerating partial order planners using a fuzzy gentics based leaning algorithm

planning is automatic genration of a sequence of actions to transfer the world from an initial to a desired final state to perform this one of the basic tasks in each planning step is the selection of an appropriate action to add to the plan under development in most cases there are several candidate actions to choose . selecting the best action in each planning step has a major effect on the planner overall performance. this paper introduces a new learning method for accelerating partial order planners the method is based on a fuzzy genetics based learning algorithm. the main idea of the algorithm ia based on sorting the cndidate actions according to their usefulness.

کلیدواژه ها:

planning ، partial order planners ، fuzzy logic ، genetic algorithm ، rule learning