Fair-efficient call admission control policies for broadband networks

An application of cooperative game theory to the synthesis of fair call admission controls for multi-service loss networks is presented. Three arbitration schemes are studied: Nash, Raiffa-Kalai-Smorodinsky and modified Thompson. The proposed model for evaluation of these schemes is based on the value iteration algorithm from Markov decision theory. The arbitration schemes are compared with two traditional call admission objectives, traffic maximization and blocking equalization. The comparison demonstrates that the arbitration solutions provide some attractive fairness features not possessed by traditional objectives, especially in overload conditions. In particular, traffic maximization can result in a total rejection of some services under heavy overload. A dynamic arbitration scheme is proposed, where the solution depends on some agreement point related to nominal conditions. In this approach, the increase of the throughput caused by the overload can be fairly distributed among all network users.<<ETX>>