A Wireless Revenue Based Scheduler with QoS Support

We propose a revenue based scheduling approach for a time slotted wireless communications system that takes Quality-of-Service (QoS) requirements such as a certain minimum data rate, maximum bit error rate, maximum delay or maximum delay variance into account. The objective function to be optimized is the revenue of the network operator. The QoS requirements are implicitly mapped onto the revenue by utilizing appropriate income functions. The resulting optimization problem is too complex to solve it in closed form. However, by introducing some simplifications it can be transformed into the problem of calculating a metric for each user in each time slot and scheduling the user who has the largest metric. Additionally, we propose an even simpler heuristic scheduler. The performance of the two approaches is evaluated by Monte Carlo simulations and compared with an upper bound of the achievable revenue. It turns out that the analytically derived scheduler achieves 99 % of this upper bound. I