Lightweight max weight scheduling algorithms for wireless networks

Abstract

We propose a class of binary queue length information based max-weight scheduling algorithms for wireless networks. In these algorithms, the scheduler, in addition to channel states, only needs to know when a link's queue length crosses a prescribed threshold. We show that these algorithms are throughput optimal. Further, we incorporate time-since-last-service (TSLS) information to improve delay and service regularity of the scheduling algorithms while ensuring throughput optimality. We also perform simulations to illustrate throughput, delay and service regularity performance of the proposed algorithms.