OPTIMAL-CONTROL OF AN M/M/S QUEUING SYSTEM

The authors consider the problem of determining optimal operating policies for an M/M/S queueing system. The system state (x//1, x//2) is defined by x//1 the number of customers in the system, and x//2 the number of active service channels. The cost structure includes customer holding and service channel operating costs as well as linear switching cost. The authors use a continuous-time Markov decision process formulation to obtain an optimal control policy of the control limit form, and present an algorithm for the finite state problem. Finite-transition and infinite-transition with discounting are considered.