Robust supplychaindesignunderuncertaindemandinagilemanufacturing

This paperconsidersasupplychaindesignproblemforanewmarketopportunitywithuncertaindemand in anagilemanufacturingsetting.Weconsidertheintegratedoptimizationoflogisticsandproduction costs associatedwiththesupplychainmembers.Theseproblemsroutinelyoccurinawidevarietyof industries includingsemiconductormanufacturing,multi-tierautomotivesupplychains,andconsumer appliances tonameafew.Therearetwotypesofdecisionvariables:binaryvariablesforselectionof companies toformthesupplychainandcontinuousvariablesassociatedwithproductionplanning.A scenario approachisusedtohandletheuncertaintyofdemand.Theformulationisarobustoptimization model withthreecomponentsintheobjectivefunction:expectedtotalcosts,costvariabilitydueto demand uncertainty,andexpectedpenaltyfordemandunmetattheendoftheplanninghorizon.The increase ofcomputationaltimewiththenumbersofechelonsandmembersperechelonnecessitatesa heuristic. Aheuristicbasedona k-shortest pathalgorithmisdevelopedbyusingasurrogatedistanceto denote theeffectivenessofeachmemberinthesupplychain.Theheuristiccanfindanoptimalsolution very quicklyinsomesmall-andmedium-sizecases.Forlargeproblems,a“good”solutionwithasmall gap relativetoourlowerboundisobtainedinashortcomputationaltime.