Determination of Energy Efficient Routing Protocol for Underwater Optical Wireless Sensor Networks Using Reinforcement Learning Algorithm

underwater optical wireless sensor networks (OWSNs) are rapidly emerging networking technologies due to their quick response and wide transmission range. The light signals in OWSNs can travel through a long distance and result in higher communication reliability and longer-range sensing. In order to support a reliable communication of these networks, efficient routing and resource management schemes, which can adapt dynamically to the changing environment, are of vital importance. Reinforcement learning (RL) provides a versatile framework to optimize the routing and security parameters of OWSNs by learning from the environment and modeling the network. This paper presents an overview of RL-based approaches for routing and security in OWSNs. Various RL strategies have been proposed to address the routing, including Q-learning, SARSA, and double Q-learning. As for security, RL can be used to identify threats, detect anomalies, and reserve resources through the simulated interaction with the environment. The advantages of using the RL framework for underwater optical wireless sensor networks are discussed in this paper.