logo True scholar network
Back to Top

Optimization in Large Scale Problems

January, 1970

Journal Name:

DOI: 10.1007/978-3-030-28565-4 move

Issue: | Volume: | Page No:

Abstract

TheconceptofIndustry4.0,smartmanufacturing,orFourthIndustrialRevolution signifiesdevelopmentandinvolvesthedigitalizationofproductionprocessesin industry.Itsprogressischaracterizedbycontinuousgrowthandrapidchanges,with thegoalofprosperingandimprovinglivingstandardsthroughhighvalue-added servicesandproducts.Ontheotherhand,focusonpeopleratherthanindustrybrings otherphenomenaintoconsiderationthatconvincegovernmentstoputsocietyatthe centerratherthanindustry,takingthetechnologyasacatalystanddriver,looking forthegeneralwelfareofcitizens,andplacingthepeopleatthecenterofIndustry 4.0.ThisviewpointofcivilizationiscalledSociety5.0. Duetothefactthatchangingtheworkingstylewillchangeourlifestyleor viceversa,therelationshipbetweenIndustry4.0andSociety5.0isundeniable.In Industry4.0,thegenerationofknowledgeandintelligenceisdonebyhumanswith thehelpoftechnology;inSociety5.0,thegenerationofknowledgeandintelligence willcomefrommachinesthroughartificialintelligenceattheserviceofpeople. Regardlessofthegeneratorofknowledge,information,anddata(eithersocietyor industry),thescaleofproblemsisbiginthisdomainsuchasageingpopulations, climatechange,ridesharing,energygridmanagement,andvehicleroutingproblem. Todealwiththesechallenges,ahugeamountofdataisneededtobecollectedand analyzedtofullyleverageitsbenefits.Theusesofanalyticalmethods(mathematical optimization,heuristicmethods,decompositionmethods,stochasticoptimization, etc.)haveafirmtrackrecordofanalyzingthevastamountsofdatatooptimizethe planningandreal-timecontrolofprocesses.Therefore,operationsresearchmethods inthecontentofoptimizationforlarge-scaleproblemsarewell-positionedtobenefit Society5.0andIndustry4.0.However,howthesolutionsmighttransitionfrom theorytorealapplicationsisaquestionthatthisbookisgoingtoanswer. Thisbookhastheaimofprovidingresourcefulthinkingandinsightfulmanagementsolutiontomanychallengesthatdecision-makersmayfaceinthepreparation, implementation,andpredictionofthefeaturesofsocietiesandindustriestoward digitalizationandsmartness.Therefore,thebookisdividedintotwosections:the firstcoveringthegeneralperspectiveandchallengesinsmartsocietyandindustry andthesecondcoveringsomecasestudiesandsolutionsfromoperationsresearch vviPreface perspectiveforlarge-scalechallengesspecifictovariousindustry-andsocietyrelatedphenomena.Chaptersfromgeneralperspectiveareauthoredbyexpert peopleineachfield,arebrief,andarevaluablesummarykeypointsforstudents, researchers,andacademicianswhoarewillingtoputfaceonthesedomainsandget ideatogoforfurtherinvestigation. Thefirstgoalofthisbookistoexplainwhatthelarge-scaleproblemsinsocioindustrialdilemmasareandwhattheoriescansupportthesechallenges.Duetothis goal,thefirstvolumeofthisseriesofworkswaspublishedwithmorefocusonthe theoryoflarge-scaleoptimizationinthebookentitledLargeScaleOptimizationin SupplyChainsandSmartManufacturing:TheoryandApplications.Lateronfrom thefeedbackofouraudience,wefoundthatbothpractitionersandacademiciansare lookingmoreforalearningplatformwhichbringsupopenquestionswithrelated casestudies.Hereuponthisneed,weeditedthecurrentbookasthesecondvolume ofthisserieswiththeaimofattentionatapplicationandwithabriefintroductionof miscellaneouschallengesinthedomainofIndustry4.0andSociety5.0. Thebookisstartedbyansweringwhylarge-scaleoptimizationisneededfor Industry4.0andSociety5.0.Followingthat,briefchaptersdiscussashortsummary oftheprincipalimplicationsofchallengesinvolvedinrunningsmartcityorindustry includingbehavioralfinanceinriskmanagementoflarge-scaleproblemsinIndustry 4.0andSociety5.0;trustingalgorithmsinSociety5.0;warehousingandmaterial handlinginIndustry4.0;product-servicesystemformanufacturingbusinessmodel transformationbydataanalytics;challengesinreliabilityengineeringproblemsin Industry4.0;theapplications,mathematics,andalgorithmsoftwo-playergame theoreticalapproachesinoptimizationproblemsforIndustry4.0andSociety5.0 applications;applicationsofqueueingtheoryinIndustry4.0andSociety5.0;online controlledexperimentsatlargescaleinSociety5.0;securitymodelinginSociety 5.0;andpedestriansimulationandtrafficmodelinginSociety5.0. Thebriefreviewofcasestudychaptersisasfollows: Chapter13presentsanindustrialmodelingandprogramminglanguage(IMPL)for optimizationmodelingandestimationofindustrialprojectssuchasoilandgas, chemicals,miningandminerals,pulpandpaper,andfoodandbeverage.Itis ahighlycomprehensivestructure-andsemantic-basedlanguageforindustrial off-andonlineoptimizationmodeling.Mostmixedintegerlinearprogramming (MILP)andnonlinearprogramming(NLP)solversareconnectedtoIMPLto solveavarietyofoptimizationproblemssuchasdesign,planning,scheduling, operations,andprocesscoordinatingoptimizationproblems. Manyalgorithmsarecategorizedas“machinelearning”suchassupportvector machine,logisticregression,graphicalmodels,anddeeplearninginIndustry 4.0andSociety5.0whichnotonlysummarizeourdatabutareperceivedas alearningmodelorclassifierfromthedataandconsequentlyfindthehidden patternindatathatwillbeseeninthefuture.Chapter14studieshowtotrain effectivelylarge-scalemachinelearningmodelsbasedonstochasticgradient method.Prefacevii Inmilitarylogistics,therearethousandsofvaluableinventoriestobemanaged. Chapter15discussesaboutawholesaleinventoryoptimizationmodelforUS Navy. Semiconductormanufacturingisacapital-intensiveindustry,inwhichmatching thedemandandcapacityisachallengingdecisionduetothelongleadtime forcapacityexpansionandshorteningproductlifecyclesofvariousdemands. Chapter16studiesthemulti-objectiveproduct-mixplanningandrevenuemanagementforthemanufacturingsystemswithunrelatedparallelmachineswitha multi-objectivegeneticalgorithm. DecompositiontechniquessuchasBendersdecompositionhaveacriticalrolein large-scaleoptimizationwithsuccessfulresultsforvariouslarge-scalemixed integerlinear,nonlinear,convex,andnon-convexoptimizationproblems.Chapter17proposesaframeworktosolvesetcoverproblems(SCPs)withblockangularstructuresthroughsolvingtheirsubproblemsandthendevelopamethod forsolvinggeneralSCPs. ManagingelectronicsupplychainsisachallengingissueinIndustry4.0,andrapid responsetocustomerordersisnecessarytodetermineaneffectivelong-term riskmitigationstrategyforthesebusinesses.Chapter18proposesarisk-based stochasticoptimizationframeworkforelectronicsupplychainsbasedonhybrid fabrication-fulfillmentmanufacturing. WehavemanylocationandinventorydecisionproblemsinSociety5.0.Thereisa challengeonoptimaldecision-makingonthesparesallocationandthebudgeting problemsinmultiple-locationinventorysystems.Chapter19analyzestheeffect ofcustomerpatienceonmultiple-locationinventorysystems. Health4.0dealswithhowvirtualizationandbigdataarerevolutionizinghealthcare inmodernsociety.Inaddition,morerobustmodelsinthepracticalhealthcare environmentaredemanded.Medicalimagingcanfacilitatediagnosis,treatment, andsurgicalplanningandincreaseclinicalproductivity.Artificialintelligence (AI)techniqueshelptheaccuracyandefficiencyofimageprocessinginhealthcareservices.Chapter20studiestwodeeplearning-basedAImethodsfor high-dimensionalmedicalimageanalysis,e.g.,tissueclassificationandmedical imagedataaugmentation. WearefacinganexponentialgrowthinusingelectricalvehiclesinSociety5.0 becauseoftheirkeyroleinsustainabledevelopment.Futuresmartcitiesaffect advancedtransportationnetworksbyglobalelectrification.Chapter21studies thecomprehensivebehaviorofelectricvehicledriversbyconsideringphysical characteristicsofelectricvehiclestoevaluatetherequiredbatterypowerfor overcomingmechanicalresistances. Complexsystemssuchassocialnetworksareanalyzedthroughnetworkmodeling. NetworkingisabigchallengeinSociety5.0.Itismergedwithinfluence maximization(IM)problemofidentifyingasmallsubsetofinfluentialpeople tomaximizetheirspreadofinfluenceinanetwork.Chapter22discussabout influencemaximizationinsocialnetworksandhasmanyapplicationssuchas viralmarketing,electioncampaign,counterterrorismefforts,rumorcontrol,and salespromotions.viiiPreface Columngenerationhasbeenusedsuccessfullytosolveavarietyoflarge-scale optimizationproblems.Chapter23discussesabouthowtobuildeffectivecolumn generationmodelstosolvereal-worldlarge-scaleoptimizationproblemswith applicationinairlineindustry. Curbspacemanagementandtrafficflowinsmartcitiesaretwoessentialelements ofthetransportationsystemthatassociatewitheachotherandaffecttheoverall systemperformance.Thegrowthofnewmobilityoperatorsandgoodsdelivery inurbanareasresultsinagrowingdemandforpickup/drop-offaccesstothe curbsides.Chapter24investigatestheallocationofcurbspaceforvarioususesto enhancetheurbanmodalitysystems’performancethatcanimprovetheoverall transportationsystemperformance. TherearemanysparseoptimizationsinIndustry4.0suchastheelectricitydemand predictionwhichiscrucialforbalancingthepowersupplyanddemandinsmart powergridsinintelligentsocieties.Chapter25presentsarealcasestudyonL1 optimizationforsparsestructuremachinelearningbasedonelectricitydemand prediction. MobilemanufacturingisanexampleofrapidtechnologydevelopmentinSociety 5.0.Chapter26presentsadynamicprogrammingmodelforanalyzingthe valueofproductioncapacitymobilityandoptimizingthelogisticsoflarge-scale productioninventorysystemswithmobileproductionunits. Millionsofsmall,family-managednanostoresaretheprimaryoriginofconsumerpackagedgoodsinmanydevelopingcountries.Chapter27dealswithamultiobjectivetravelingsalesmanproblemforderivinganeffectivesupplystrategy forreal-lifenanostoresinmajorcities. Wehopethatthisbookwillproveusefultoresearchers,students,andengineersin differentdomainswhoencounterintheirworklarge-scaleoptimizationproblems andencouragethemtoundertakeresearchinthisexcitingandpractically importantfield. Wewanttothankalltheauthorsinvolvedinthisprojectfortheircontributions.We alsowanttothankthereviewers,whohavehelpedusinreviewingandimproving severalchaptersofthisbook.