Analyzing thebehaviorsofvirtualcells(VCs)andtraditionalmanufacturingsystems: Ant colonyoptimization(ACO)-basedmetamodels

The aimofthispaperistwofold.Firstweinvestigatethethreedifferenttypesofsystems,namelycellular layout (CL),processlayout(PL)andvirtualcells(VCs).VCsareaddressedbyusingfamily-basedschedul- ing rule,developedbyapartallocationalgorithminaPL.Simulationisusedtocomparethreetypesof systems undertheperformancemetricssuchasmeanflowtimeandmeantardiness.Resultsindicate that VCshavebetterresponsivenessintermsoftheperformancemetrics.Secondwedevelopanew ant colonyoptimization-basedmetamodelsfedbyexistingsimulationrunstorepresenttheprospective simulation runs,whichrequirealotoftimeandeffort.Regressionmetamodels,whichallowustoobtain much fasterresults,areseentobepromisingtoestimatethesystemsbehaviors.