A robusttransportationsignalcontrolproblemaccountingfortrafficdynamics

Transportation systemanalysismustrelyonpredictionsofthefuturethat,bytheirverynature,contain substantial uncertainty.Futuredemand,demographics,andnetworkcapacitiesareonlyafewofthe parameters thatmustbeaccountedforinboththeplanningandeverydayoperationsoftransportation networks. Whilemanyrepercussionsofuncertaintyexist,aprimaryconcernintrafficoperationsistode- velop efficienttrafficsignaldesignsthatsatisfycertainmeasuresofshorttermfuturesystemperformance while accountingforthedifferentpossiblerealizationsoftrafficstate.Asaresult,uncertaintyhastobe incorporated inthedesignoftrafficsignalsystems.Currentdynamictrafficequilibriummodelsaccounting for signaldesign,however,arenotsuitableforquantifyingnetworkperformanceovertherangeofpossible scenarios andinanalyzingtherobustperformanceofthesystem.Thepurposeofthispaperistopro- pose anewapproach—robustsystemoptimalsignalcontrolmodel;asupply-sidewithindayoperational transportation modelwherefuturetransportationdemandisassumedtobeuncertain.Arobustdynamic system optimalmodelwithanembeddedcelltransmissionmodelisformulated.Numericalanalysisare performed onatestnetworktoillustratethebenefitsofaccountingforuncertaintyandrobustness.