N-LEVELS HEURISTIC ALGORITHM FOR PLANNING PROBLEMS

In this work we present a method using heuristic functions to improve the quality and speed of solutions for planning problems. These heuristic functions use an n-levels search prior to decide which action is better. Traditional model are implemented using only one search level to extract a quality solution plan for planning problems. Experimental results show that the quality of the solution obtained is better in many cases, when considering more levels of heuristic search instead of only one level. Additionally, the execution time of the solution plan when using an n-level search is less than the one that is obtained when using only one level. Our method also improves the performance for a variety of planning problems.