Performance Optimization in Communication Systems Using Deep Reinforcement Learning with Elite Reverse Learning Strategy - Arctic Puffin Optimization

The communication system involved in the various networking it becomes decentralized and automatic, network articles is to make decisions for increasing network performance in uncertainty network environment. The challenges of handle the large-scale network it making complex, state and action spaces are generally huge and reinforcement learning unable to identify optimum policy in reasonable time. In this research, Deep Reinforcement Learning with Elite Reverse Learning Strategy - Arctic Puffin Optimization (DRL with ERL-APO) is developed for the effective performance optimization in communication. The strategy of reverse learning strategy is incorporated with traditional APO algorithm which improves the convergence rate of traditional APO algorithm for efficient communication performance optimization. For ensuring the high rate for less channel gained the users who needs much energy in terms of deprived energy share. Therefore, outcomes on substantial minimize in sum rate. The proposed DRL-APO with ELS obtained execution time of 8 on 10 SNR, 9 of 20 SNR, 10 of 30 SNR, 11 of 40 SNR and 12 of 50 SNR which is better while compared with existing algorithms.