A simulatedannealingalgorithmforsinglemachineschedulingproblemswithfamily setups

Motivated bythereal-lifeschedulingprobleminasteel-wirefactoryinChina,thispaperstudiesthe problem ofminimizingthemaximumlatenessonasinglemachinewithfamilysetups.Inviewofthe NP-hard natureoftheproblem,neighborhoodpropertiesoftheproblemareinvestigated.Itisfound that thetraditionalmove-basedneighborhoodisinefficienttosearch.Thenanewneighborhood,which is basedonbatchdestructionandconstruction,isdeveloped.Asimulatedannealingalgorithmwiththe new neighborhoodisproposed.Experimentsarecarriedoutontherandomlygeneratedproblemsand the real-lifeinstancesfromafactoryinChina.Computationalresultsshowthattheproposedalgorithm can obtainbetternearoptimalsolutionsthantheexistingalgorithm.