A DSSforShift Design andWorkforce Allocation ina CallCenter

Workforce scheduling, shift designand staff allocation arethemajorproblems incall center management. The aim ofthecallcentermanageristo allocate and dynamically update theworkforce sothattheincoming calls are answered intheshortest possible time, abovecertain service level measures. Thesoftware tools developed toaiddecision makinginthese areasusethemodelsthatarebasedonErlangC calculations. Howeverthestrict assumptions ofErlang-C often leadtoinvalid decisions. Especially atpeaktimesduring the day,dynamic updates intheproposed design isinevitable. In thisstudy, a frameworkfora decision supportsystemis proposedfordesigning theshifts andallocating theagent workforce totheshifts ina callcenter, sothattarget service levels aremet.Intheproposed system, shifts aredesigned by solving alinear optimization model.Usingthissolution asthe input, asimulation modelisdeveloped todynamically update theworkforce design sothattheminimumrequired service level ismetatalltimes. Theproposed DSSisillustrated onanexisting callcentersystem,alternative designs aregenerated and compared. approximations. Intoday's complex andmulti skilled call centers, thestrict assumptions ofErlang C areoften invalid andinseveral cases theresulting solutions aremisleading. Actually itisverydifficult touseErlang modelforthe"what if'scenarios generated toevaluate call center design, call handling strategies andcall routing options. Thustheuseof analternative tool becomes inevitable andwithitsdynamic anddeveloping nature, system simulation appears tobean appropriate methodtobeusedinwhat-if analysis in workforce management. Inthisstudy, a decision support system(DSS)is proposed fordesigning theshifts andallocating theagent workforce totheshifts inacall center sothatthetarget service levels aremetinallperiods. Intheproposed system, shifts aredesigned bysolving alinear optimization model. Usingthissolution astheinput, a simulation modelis developed todynamically update theworkforce allocation so that theminimumrequired service levels aremetatall times. Oneoftheleading call center companies ofTurkey has beenchosen forthis study. Currently, thecall center provides services inthree different locations withdifferent agent and seatcapacities. A majorproblem istoallocate thecalls between these locations, while attaining thetarget service and utilization levels. At present, shift scheduling andstaff allocation decisions aregiven bytheuseofaworkforce management tool. However, theystill haveproblems in reaching thetarget service levels atpeakhours, mostly dueto thereasons explained above. Increasing thenumber ofagents allocated totheshifts results inseverely lowutilizations in theperiods wherethecall rates arelow.Generating shifts withflexible durations appears tobeareasonable option for this problem. Theaimofthis study istodevelop aneasyto useenvironment byARENA andMS-Excel that will work together withtheexisting toolandenhance itssolution quality bygenerating valid results andbyallowing real time assessment ofnumerousalternatives on flexible shift durations inmultiple locations. As anillustration, an alternative workforce schedule andstaff allocation is designed byusing theproposed system anditisshownthat theproposed design outperforms thesolution ofthecurrent system. Theorganization ofthepaper isasfollows: Insection II, a literature review on thebasicmodelsdeveloped for workforce scheduling andstaffing isprovided. Insection III, aDSSisdesigned forworkforce management that integrates an LP modelanda simulation model. InSection IV, alternative workforce designs aregenerated andcompared withthesolution ofthecurrent system. Insection V,the framework oftheDSSisdemonstrated. Section VIincludes theconclusion andthesummary.