Solve The Problem With Precedence Constraints By Combining Multiple Heuristic

Nowadays, Intelligent Planning is one of the most significant research fields in AI.The research on Intelligent Planning derived from 1960s.Because of the new breakthroughs in problem solving and problem description, Intelligent Planning has become a popular trend. Heuristic search solve the problem via combining knowledge and vast experience,always can solve problem efficiently.Fast-Forward planning system is exactly a wonderful state-space search planner,which has shown excellent performance in most STRIPS domains.There always appears that when we achieve a sub-goal,it will delete another subgoal which achieved before it ,and the removed subgoals should be realized once again in the planning,so it is a waste of time and space. In our papers,we develop a new algorithm to solve with the precedence constraints.Assuming that the subgoals are much enough,first, we divide the sub-goals into one group which have a special relationship with each other,so it is independence between groups;Sencond,We sort the subgoals within the group;the third,testing them with function ctx within group,then among groups;and finally,we develop one heuristic based on h^cea,then achieve our goals with some heuristics run in parallel.Because of algorithm dealing with most work during the preprocessortime,so apply combining the heuristic h^cemwith heuristic h^pccis superior to apply the heuristic h^ceafor speed;although the information carried by h^cemis subtertminal to that of heuristic h^cea,it guarantee the admissibility,which h^ceacan't.So this heuristic search on AI planning is important both theoretically and practically.