Exploiting wireless channel state information for throughput maximization

The problem of scheduling packets over a number of channels with time varying connectivity is considered. Policies proposed for this problem either stabilize the system when the arrival rates are within the stability region, or optimize an objective function under the assumption that all channel queues are saturated. We address the realistic situation where it is not known a priori whether the channel queues are saturated or not, and provide a scheduling policy that maximizes the weighted sum of channel throughputs. We employ a burstiness-constrained channel model that allows us to dispense of statistical assumptions and simplifies the proofs.