Fuzzy admission control of customers to a three-workstation queueing network

We consider a queueing network with two classes of customers and three workstations, each consists of one server and one buffer. One workstation is in the first stage and two parallel workstations in the second stage. Each workstation in the second stage receives customers from its own arrival stream. The buffer in the first stage consists of sufficient customers of both classes. A reward is earned whenever the first stage accepts a customer and a holding cost is incurred by a customer in the queue in the second stage. The objective is to dynamically select the class of customer to be next served at the first stage, based on the state of the system, so as to maximize the average benefit over an infinite horizon. In the paper, a fuzzy approach is presented to solve this complex problem. Simulation shows that the approach is efficient and promising.