Optimality of Certain Channel Aware Scheduling Policies

Abstract

With an abstraction of serving rate-adaptive sources on a broadcast-type wireless channel as a utility maximization problem, it is shown how one can design many intuitive online scheduling policies based upon the feedback that one obtains at the scheduler. Using a stochastic approximation argument it is then shown that the constructed algorithms converge to optimal solutions of the utility maximization problem over different sets which critically depend on the quality of the feedback information. 1