Network utility maximization in multipath lossy wireless networks

Abstract

Summary The end‐to‐end rate control in multipath wireless multihop networks has been widely studied so as to exploit path diversity and improve network performance. However, it has been recognized that the hop‐by‐hop congestion control is stable and spatially smoothed in the absence and presence of delay, respectively. As a result, we are motivated to consider the lossy features of wireless links and propose a novel utility maximization problem for multipath communication networks. In view of the fact that the considered optimization problem formulated as a multipath network utility maximization problem is in non‐convex form and hard to tackle, we first cast the underlying problem into a convex one and then solve the convex approximation problem by the Lagrangian duality technique. What is more, an efficient hop‐by‐hop rate control algorithm is proposed. In addition, we prove that the proposed two‐loop algorithm converges to the optimal solution satisfying the Karush–Kuhn–Tucker optimality conditions of the original optimization problem. We finally validate the effectiveness of the proposed algorithm through extensive simulations and compare the performance of the proposed algorithm with a hop‐by‐hop single‐path rate control algorithm and also with an end‐to‐end multipath algorithm. Copyright © 2015 John Wiley & Sons, Ltd.