Distributed Joint Resource Allocation in Multi-Radio Multi-Channel Wireless Mesh Networks

This paper studies how to design distributed resource allocation in multi-radio multi-channel wireless mesh networks with the objective of maximizing the network utility. We address the problem via a cross-layer approach with a joint consideration of multi-path routing, congestion control, scheduling, radio allocation and channel assignment. The problem is formulated as a mixed-integer non-linear programming (MINLP), which normally requires a centralized solution and hence has prohibitively high computation complexity. We propose a two-phase distributed mechanism to substantially reduce the computation load and communications overhead. Firstly, we relax the integral variables and obtain a convex programming, which serves as the upper bound of the optimal utility. Secondly, we propose a distributed scheme to approach the upper bound within the feasible region of the optimization problem. To evaluate the performance of the distributed algorithm, we compare it to the exact optimal solution to the MINLP objective function, which is obtained by using a centralized branch-and-bound method. Simulation results show that the performance of our proposed distributed algorithm is close to the optimal solution.