Algorithm For Staffing And Scheduling Problem In Time-varying Call Service

With the rapid aggravation of intense competitiveness in the market, modern enterprises need to provide not only customized products but also satisfactory service. The diverse and stochastic demands from customer, the changing trend from product competition to service competition between companies highlight the importance of both quality and efficiency, which is the focus of manager and academia as well. Recently, call centre becomes the prevalent style to provide service to customers in various industries. The cost of human resources accounts for more than half of the operation cost in most of the call centers, which explain why the research about how to effectively schedule the staff receives so much attention.Relying on national natural science foundation, the thesis did some research on staffing and scheduling problem emphasized in call service operation management. Having found the shortcomings in the traditional two-stage method, this paper used improved simulation-based analysis centre cutting-plane method to solve joint staffing and scheduling problem under single-skill and time-varying condition. The main work consist of three aspects:First, through ProModel simulation, this paper pointed out solving the staffing and scheduling problems separately caused a waste of personnel leading to high operational cost in together with neglecting the interaction between adjacent periods. Second, this paper built the model for joint staffing and scheduling problem using average speed of answering as measure for service quality. In addition, this paper demonstrated the rationality of solving sample-average approximate instead of the original stochastic optimization problem. Third, this paper used improved cutting-plane method to solve the sample-average approximate problem, in the process of which this paper modified the generation of discrete pseudogradient and the addition of constraints when the discrete pseudogradient equals zero. Moreover, this paper proved the convergence of the improved algorithm. At last, this paper verified the effectiveness by solving an example using AMPL modeling language with MINOS and CPLEX solver.Utilizing simulation for stochastic problems is a promising aspect for further study. The simulation method of calculating service level this paper used is also applicable to other performance which is hard to represent by mathematic expression. Although the results of this paper is obtained under some theoretic hypothesis, it is of practical value for staffing and scheduling in call center with high personnel cost.