Estimation-basedmetaheuristicsfortheprobabilistictraveling salesman problem

The probabilistictravelingsalesmanproblem (PTSP)isacentralprobleminstochasticrouting.Recently, we haveshownthatempiricalestimationisapromisingapproachtodevisehighlyeffectivelocalsearch algorithmsforthePTSP.Inthispaper,wecustomizetwometaheuristics,aniteratedlocalsearch algorithmandamemeticalgorithm,tosolvethePTSP.Thiscustomizationconsistsinadoptingthe estimationapproachtoevaluatethesolutioncost,exploitingarecentlydevelopedestimation-based local searchalgorithm,andtuningthemetaheuristicsparameters.Wepresentanexperimentalstudyof the estimation-basedmetaheuristicalgorithmsonanumberofinstanceclasses.Theresultsshowthat the proposedalgorithmsarehighlyeffectiveandthattheydefineanewstate-of-the-artforthePTSP.