WLC39-4: Joint Power Allocation and Scheduling for Deadline Constrained Wireless Traffic

We study the problem of joint power allocation and scheduling for deadline constrained traffic on the downlink of a cellular wireless communication system, where multiple users can be scheduled simultaneously on orthogonal spreading codes (HSDPA), or sub-carriers (OFDMA), subject to a sum power constraint. We formulate the problem for a canonical two-user model within a dynamic programming (DP) framework. We present key structural properties of the optimal solution to the DP, and leverage the intuition thus gained into the fundamental power allocation and scheduling trade-offs to propose the Channel-Aware Power Allocation and Scheduling (CAPAS) policy. We demonstrate the efficacy of CAPAS relative to benchmark schedulers via simulations. The key conclusion from our work is that scheduling one user at a time is sub- optimal for deadline constrained wireless traffic, both in terms of number of missed packet deadlines, and average transmit power consumption at the base-station.