RELAY SELECTION SCHEME BASED ON DEEP REINFORCEMENT LEARNING IN INTERNET OF UNDERWATER THINGS

To address the issues of limited communication coverage and energy resources in the Underwater Internet of Things (IoUT), a relay selection strategy is proposed that combines optimal transmission power with proximal policy optimization (PPO). The strategy established a Markov model for the underwater cooperative relay process, and employed convex optimization to determine the optimal transmission power for the source node and the selected underwater relay node, maximizing the signal- to- noise ratio. Simulation results show that under the same conditions, compared with random relay selection and Q- learning (QL) algorithm, this approach has higher cumulative rewards, a higher signal- to- noise ratio, and lower interruption probability.