A branch-and-pricealgorithmforanintegratedproductionand inventory routingproblem

With globalization,theneedtobetterintegrateproductionanddistributiondecisionshasbecomeever more pressingformanufacturerstryingtostreamlinetheirsupplychain.Thispaperinvestigatesa previouslydevelopedmixed-integerprogramming(MIP)modelaimedatminimizingproduction, inventory,anddeliverycostsacrossthevariousstagesofthesystem.Theproblembeingmodeled includesasingleproductionfacility,asetofcustomerswithtimevaryingdemand,afiniteplanning horizon, andafleetofhomogeneousvehicles.Demandcanbesatisfiedfromeitherinventoryheldata customersiteorfromdailyproductdistribution.Whetheracustomerisvisitedonaparticulardayis determinedbyanimplicittradeoffbetweeninventoryanddistributioncosts.Oncethedecisionismade, a vehicleroutingproblemmustbesolvedforthosecustomerswhoarescheduledforadelivery. A hybridmethodologythatcombinesexactandheuristicprocedureswithinabranch-and-price frameworkisdevelopedtosolvetheunderlyingMIP.Theapproachtakesadvantageoftheefficiencyof heuristicsandtheprecisionofbranchandprice.Implementationrequireddevisinganewbranching strategy toaccommodatetheuniquedegeneracycharacteristicsofthemasterproblem,andanew procedureforhandlingsymmetry.Anovelcolumngenerationheuristicandaroundingheuristicwere also implementedtoimprovealgorithmicefficiency.Computationaltestingonstandarddatasets showedthatthehybridschemecansolveinstanceswithupto50customersand8timeperiodswithin 1 h.ThislevelofperformancecouldnotbematchedbyeitherCPLEXorstandardbranchandpricealone.