| We propose an admission and routing control policy for a network of service facilities in a stochastic setting in order to maximize the expected reward rate for the policymaker. The policy is a target tracking policy: a linear program provides a target operating point and an exponential penalty function is used to translate the optimal deterministic point into a feasible admission policy. For single service facility loss systems, closed form expressions for mean and variance of steady state reward rate are given and the policy is shown to be superior to an independent thinning control policy. When queueing and reneging are allowed, a lower bound on the expected reward rate is derived and we prove that the policy performs well both in the transient period as well as in steady state as long as the capacity request sizes are relatively small compared to the capacity of the service facility. Simulation studies are included to illustrate that our policy also performs well when request sizes are moderate or large with respect to the capacity. We present a (heuristically derived) diffusion approximation for a single service facility under the control policy when queueing and reneging are allowed. The results are extended to more general feedforward networks. |
