Packet scheduling with revenue optimization and weighted delay minimization

In this paper we propose an adaptive packet scheduling model for networks providing quality of service. The model minimizes the weighted mean delays of the connections for specified pricing classes (gold, silver, and bronze), while also maximizing the revenue of the service provider. In this scenario a call admission control (CAC) mechanism is used for guaranteeing specified mean delay characteristics for the different service classes. A closed form analytic and optimal solution is derived for the weights of the scheduler. The scenario is independent of the statistical assumptions for the connections, and is therefore robust against erroneous estimates of the traffic characteristics