5 CONSTRAINT SATISFACTION PROBLEMS

← Back to document page

5CONSTRAINTSATISFACTIONPROBLEMSInwhich weseehowtreatingstatesasmore thanjustlittleblack boxesleadsto theinventionofa range ofpowerfulnew search methodsanda deeperunderstandingofproblemstructure and4 exploredtheideathatproblemscanbesolvedby searchinginaspaceofstates. Thesestatescanbeevaluatedbydomain-specif icheuristicsandtestedtoseewhetherthey ,however,eachstateis is representedbyanarbi-BLACKBOXtrarydatastr ucturethatcanbeaccessedonlybytheproblem- specificroutines thesuccessorfunction,heuristicfunction, , whosestatesandgoaltestconformtoa standard,structured,andverysimplereprese ntation( ).Searchal-REPRESENTATIONgorithmscanbede finedthattake advantageofthestructureofstatesandusegen eral-purposeratherthanproblem-specifiche uristicsto enablethesolutionoflargeproblems( ).Perhapsmostimportantly, thestandardrepresentationofthegoaltestre vealsthestruc-tureoftheproblemitself( ).

rather than problem-specific heuristics to enable the solution of large problems (Sections 5.2– 5.3). Perhaps most importantly, the standard representation of the goal test reveals the struc-ture of the problem itself (Section 5.4). This leads to methods for problem decomposition

  Tests, Satisfaction, Section, Constraints, Constraint satisfaction