A Comprehensive Review of Reinforcement Learning Based Routing Algorithms for Wireless Sensor Networks

Wireless Sensor Networks (WSNs) are commonly utilized in remote environments.Intelligent sensors are embedded to monitor numerous parameters, including pressure and temperature.The sensors, known as nodes, gather data and transmit it to the base station.Nodes in WSNs must acquire information from their environment due to their dynamic nature.In WSN, energy economy and network longevity are essential goals that routing can facilitate.The changing nature of WSNs complicates adaptation of conventional routing approaches and computational intelligence systems.Reinforcement Learning (RL), a subset of computational intelligence approaches, facilitates the selection of future actions based on insights acquired from previous encounters with the environment.Consequently, RL is often employed to develop routing algorithms in WSNs.Recent research indicates that employing a reinforcement learning-based Q-learning technique can yield the most energy-efficient pathways in WSNs.The document presents a comprehensive examination of routing methodologies utilizing RL in WSNs.The Q-learning strategy is extensively discussed as it underpins other RL strategies.The paper elaborates extensively on the selection of state-action pairs and rewards in Q-learning across various literary works.The experimental study presents optimization policies and challenges for enhancing Quality of Service (QoS).