Optimization of wireless multi-hop networks with random access

Abstract

Cross-layer design of medium access, routing and congestion control is considered for multi-hop wireless ad hoc networks with random access. A network utility maximization frame-work is adopted to jointly optimize link persistence probabilities, per-link flow rates, and end-to-end throughput. The optimization problem with flow conservation and probabilistic random access constraints is inherently non-convex. A series of convex surrogate problems is thus formulated using a successive convex approximation technique. A distributed algorithm is then developed based on the dual method, which is convergent to a (locally) optimal solution. To much practical appeal, solutions to the subproblems corresponding to different layers of the protocol stack, are obtained at each node with very low computational complexity. Numerical tests verify the proposed novel design.