Convergence rate control for distributed multi-hop wireless mesh networks

Abstract

Control message delay is a major factor affecting distributed network utility maximization (NUM) in practical wireless mesh networks, and the inherent delay of the control response leads to severe oscillations and slow convergence. However, existing NUM techniques either fail to pay attention to the convergence rate or achieve suboptimal results. This paper proposes a novel technique to solve the reverse impact of inherent control delay in the distributed NUM of wireless mesh networks with noisy feedback by overcoming the slow convergence induced by control delay in conjunction with feedback noise. A self-tuning algorithm reduces the number of required iterations for convergence while also providing an optimum value in distributed networks, and substantive simulations prove this improved convergence performance within a distributed multi-hop wireless mesh network. A theoretical explanation built from the viewpoint of control theory is also put forward to describe the model.