Optimal Call Admission Control Policy In Wireless Networks

It is well known that the call admission policy has a big impact on the performance of a wireless network. Due to the nonlinear dependence between new calls and handoff calls, it becomes quite difficult to exactly model a wireless network. In this paper, we take all the incoming calls to a single cell of a wireless network as one class, their distribution as a general distribution and model it as a semi-Markov decision process (SMDP). Assuming that accepting each call would contribute a reward to the system and the system incurs a cost per unit time for the calls waiting in the buffer, we derive the optimal call admission control policy for the maximum reward. By using SMDP technique, we find that optimal call admission control policy is a control limit policy if the waiting cost function is a convex nondecreasing function.