Optimal Control Policy for a Queuing Service System

MA Yong-kai
2012-01-01
Abstract:This paper builds a double-threshold queuing service model for coordinating customer’s waiting time and operating costs.Using an elaborated probability decomposition technique,we derive the transient and stationary queuelength probability distribution.Next,according to the system steady-state performance index,the system net profit function is modeled.A global algorithmic approach to optimize the dual discrete variable function is proposed,which is for learning the optimal control policy.Furthermore,numerical experiments show the advantages of the double threshold control policy.
What problem does this paper attempt to address?