Asynchronous Distributed Joint Optimization in Wireless Multi-Hop Networks

In this work, we resolve a cross-layer distributed optimization problem in wireless multi-hop networks that can jointly maximize network lifetime and optimize system utility. For this problem, even a synchronous distributed algorithm would be unacceptable due to simultaneous computations required and a global order of transmission to be known in advance. Therefore, we derive an edge-based formulation for consensus agreement on the variables involved, and develop an asynchronous decentralized algorithm specific to the joint optimization problem based on alternating direction method of multipliers (ADMM). Our numerical results show that the ADMM-based algorithm can make a good trade-off between the heterogeneous optimization objectives with excellent computational efficiency, and exhibit that the algorithm can converge faster to a stable solution than a gossip-based algorithm for the same aim, and as efficiently as the synchronous counterpart.