Randomized Algorithms for Throughput-Optimality and Fairness in Wireless Networks

In this paper, we study multi-hop wireless networks with general interference models, and describe a practical randomized routing-scheduling-congestion-control mechanism that is guaranteed to fully utilize the network capacity, and achieve fair allocation of the resources. Using the framework described in this paper, low complexity distributed algorithms can be developed for a large class of interference models and fairness criteria (such as proportional and max-min fairness). Earlier distributed algorithms proposed in this context have been highly interference model dependent, and can guarantee at most 50% utilization of the achievable throughput. This is the first work that assures 100% utilization and also provides fair allocation in multi-hop wireless networks