A Hybrid Of Search Efficiency Mechanisms: Pruning Learning Heuristic Hybrid (Plh2)

This paper introduces hybrids as the major means of incorporating proper intelligence into a systematic search. A Pruning Learning Heuristic Hybrid (PLH2) is presented that uses a set of accelerating data structures and exploits the synergy existing between heuristic methods, learning mechanisms, and pruning rules. This hybrid can easily use abstraction mechanisms. In PLH2, using effective data structures, a heuristic method is combined with pruning rules that together facilitate an efficient learning mechanism. The effective combination of pruning rules with a simple learning mechanism makes the heuristic method more powerful and reduces the search effort. Partial updating of the contents of facilitating data structures associated with the nodes reduces the search effort significantly. The computational experiments show that PLH2 efficiently tackles a practical problem of wide generality and the performance analysis indicates PLH2 can present several advantages over A* under specific conditions.