Dynamic Scheduling with Statistical Delay Guarantees and Traffic Dropping

This work studies the dynamic scheduling problems in wireless networks with delay-sensitive loss-tolerant users. The users' traffic satisfies some statistical delay constraints. Moreover, the traffic can be dropped but the dropping rates do not exceed some thresholds. We consider two scheduling scenarios. First, we study the problem to minimize the total transmission power while maintaining the minimum rates for the users. Then, we study the problem to maximize the minimum rate(s) of the users while constraining the maximum total power. We derive the optimal solutions for both scheduling problems. When the fading statistics are available, using the dual-gradient method, the optimal policies can be computed. When the fading statistics are unknown, this work proposes online scheduling algorithms using online time-averaging. The convergence and optimality of the proposed algorithm are guaranteed by the results in stochastic approximation theory.