12.02.2014 Views

Disjoint Pattern Database Heuristics

Disjoint Pattern Database Heuristics

Disjoint Pattern Database Heuristics

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

tiles,wewanttogrouptogethertilesthatareneareachotherinthegoal partoftheindextothedatabase.Asageneralrule,whenpartitioningthe entries,whichrangefrom0to38moves.Inneithercaseistheblankposition rangefrom0to33moves.Theeight-tiledatabasecontains518;918;400<br />

state,sincethesetileswillinteractthemostwithoneanother.<br />

Figure4:<strong>Disjoint</strong><strong>Database</strong>forFifteenPuzzleanditsReection<br />

4.1 4 ExperimentalResults<br />

Wefoundalloptimalsolutionsto1000randomFifteenPuzzleprobleminstances,usingIDA*withavarietyofheuristics.Theaverageoptimalsolution<br />

lengthoftheseinstancesis52.522moves,andtheaveragenumberofoptimal<br />

instancetondtherstoptimalsolution.Thethirdcolumndisplaysthe solutionsis15.9.Table1showstheresults.Therstdatacolumnshows<br />

averagespeedofthealgorithm,innodespersecond,ona440MegaHertz secondcolumngivestheaveragenumberofnodesgeneratedperproblem theaveragevalueoftheheuristicfunctionoverthe1000initialstates.The<br />

instance. averagenumberofnodesgeneratedtondalloptimalsolutionstoaproblem time,inseconds,tondtherstoptimalsolution.Thelastcolumngivesthe SunUltra10workstation.Thefourthcolumnindicatestheaveragerunning<br />

cally,thelinear-conictheuristicwastherstsignicantimprovementover Manhattandistance[5].Itappliestotilesintheirgoalroworcolumn,but secondrowisforManhattandistanceenhancedbylinearconicts.Histori-<br />

TherstrowgivesresultsfortheManhattandistanceheuristic.The<br />

statecontainsthetiles(21)inthatorder,butinthegoalstatetheyappear reversedrelativetoeachother.Forexample,assumethetoprowofagiven 10

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!