Building Large Compressed PDBs for the Sliding Tile Puzzle

Abstract

AbstractThe performance of heuristic search algorithms depends crucially on the effectiveness of the heuristic. A pattern database (PDB) is a powerful heuristic in the form of a pre-computed lookup table. Larger PDBs provide better bounds and thus allow more cut-offs in the search process. We computed 9-9-6, 9-8-7, and 8-8-8 PDBs for the 24-puzzle that are three orders of magnitude larger (up to 1.4 TB) than the 6-6-6-6 PDB. This was possible by performing a parallel breadth-first search in the compressed pattern space. Our experiments indicate an average 8-fold improvement of the 9-9-6 PDB over the 6-6-6-6 PDB on the 24-puzzle. Combining several large PDBs yields a 13-fold improvement.KeywordsHash FunctionParallel AlgorithmMemory ConsumptionSequential AlgorithmBinary Decision DiagramThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.