The Fast Forward Intelligent Planning With Backtracks And Helpful Action Ordering

Now Artificial Intelligence Planning is a very hot branch in AI. Because of its wide application, researchers pay much attention to planning technology. The topics of Artificial Intelligence Plaing focused on the speed of problem-solving and the complex modelling of domains that intelligent planning systems can deal with. Most Artificial Intelligent Planning algorithms are based on planning graph or state space search. FF based on both planning graph and heuristic state space search and performed standout in the 2nd and 3rd International Planning Competition. Because of inablity of the handling dead end state, FF does not perform ideally in some planning problem domains such as Gripper.In this paper, a new searching method enforced hill-climbing with backtracks is used, and a new method of helpful action ordering is presented. Enforced hill-climbing with backtracks can use the last searching results when enforced hill-climbing fails, so it is more fit for FF. Helpful action ordering can save the time used by enforce hill-climbing through ordering the helpful actions. We put forward a new Intelligent planner BTFF based on the enforced hill-climbing and helpful action ordering.In this paper, we investigate how to improve the enforce hill-climbing and helpful action ordering in Artificial Intelligent Planning. This research is important both theoretically and practically.