Based On Data-Driven And Robust Discrete Optimization Approaches To Staffing Problems For Call Centers

With market competition becoming more and more fierce,modern enterprises should not only provide high-quality customized products,but also provide more satisfactory service.The diversity and randomness of customer needs,and the competition between enterprises being transformed from product competition to service competition gradually,make the research quality and efficient service mode become the focus of the enterprise management as well as academia.Call center has increasingly become the common of all walks of life to provide customer service,in majority of call center,the cost of human resources accounts for more than 60%?70%for the whole operating costs.Therefore,it is particularly significant to study effective personnel work plan.This paper is based on the national natural science fund project.It researches around the staffing problems,which is the key problems in the study of operation management.By finding the deficiency of traditional method for study of uncertain problems,data-driven and robust discrete optimization approaches are proposed to improve the staffing problems,and eventually to make the staffing reduces the impact of random customer arrival.Moreover,all the models in this paper are transformed into tractable and linear counterparts,and we used IBM ILOG CPLEX 12.4 to solve them in Windows 7.The main work of this paper includes the following four aspects:the first one is modeling the staffing model according to the service flow of call centers,the result shows that the model can only solve the problem of certain,while for the uncertainty customer arrival in an actual environment,this staffing will make a huge cost loss.Therefore,it is very necessary to take into account the uncertain arrival while staffing.The second one is modeling data-driven staffing model with risk aversion,by a lot of history arrival data.Numerical experimental results verify the feasibility of data-driven model,and this staffing with optimal risk trimming factor reduces the impact of the uncertainty of arrival successfully,improves the stability of the system.The third one is using robust discrete optimization theory to model a full period robust discrete optimization staffing model for call centers.Numerical experimental results verify the feasibility of model,and this staffing with optimal robust parameter reduces the system cost and improves the deficiency of the data-driven approach,such as staffing too many agents.The last one is modeling the local period robust discrete optimization staffing model,according to the characteristics of the uncertain arrival period in actual call centers,which is improved from full period robust model.Numerical experimental results verify the feasibility of local period robust model,and this staffing with optimal robust parameter reduces the system cost as well as improves the stability of the system.The results of this paper will reduce the impact of random customer arrival and provide a comprehensively theoretical reference and support to an actual operation of call center.In addition,duing that the management of call centers operation belongs to a service field,the approaches and thought of the paper are easy to be transplanted to other service field.