A Utility-Based Approach to Bandwidth Allocation and Link Scheduling in Wireless Networks

Abstract

We study the problem of optimizing aggregate user utility in wireless ad-hoc networks under the constraints of wireless interference. We develop a market-oriented approach to bandwidth allocation with a tˆatonnement process and demonstrate its ability to effectively price bottleneck resource. One novelty is that we choose to price “interference goods” to capture the externality imposed by one application’s use of the network on other applications. In making progress we also propose a modification to the CSMA protocol that is robust enough to handle a non-schedulable bandwidth schedule. Experimental results on simulated network topologies show that the market-based approach has better scalability than alternate approximation methods and is much more efficient in terms of runtime.