Intelligence in wireless network routing through reinforcement learning

Abstract

One of the critical challenges in mobile wireless network is resource optimised routing of messages without compromising the performance criteria of the network. Routing in wireless networks has been extensively studied and a variety of routing protocols have been proposed. But these protocols experiences problems due to the dynamism in network topology. To address these problems, reinforcement learning approaches are integrated in routing solutions. This review focuses on the impact of reinforcement learning algorithms to achieve intelligence in wireless network routing. We provide contexts and benefits of applying reinforcement learning paradigm and discuss the major techniques applied to optimise routing solutions. A survey of state of the art reinforcement learning-based routing protocols is presented and categorised these protocols according to the learning strategies. We also provide open issues and suggestions for future research in improving routing solutions.