Energy-Efficient Resource Allocation in 5G Wireless Networks using Asynchronous Advantage Actor-Critic based Reinforcement Learning

5G wireless network needs a significant method to efficiently handle and allocate the resources. Exact to the network organization, the multicast communication is focussed in an effectiveness of data storage and data-based network interface. However, interface management in resource allocation is the most significant challenge in device-to-device communication. Hence, this research proposes the Asynchronous Advantage Actor-Critic (A3C) based Reinforcement Learning (RL) approach for the energy-efficient resource allocation. A3C enables the various agents to process and learn self-sufficiently in various conditions simultaneously. This asynchronous learning speeds up the convergence process, which is important in energy-efficient resource allocation where real-time decisions are required. An effectiveness of the proposed A3C-RL approach is estimated by using various performance metrices like throughput, spectral efficiency as well as energy efficiency. The experimental results demonstrates that the proposed A3C-RL approach attains the superior spectral efficiency of 99.22bits/Hz and energy efficiency of 30.39 bits/Hz on the number of SNR is fixed to 10dB as compared to Deep Q Networks (DQN) and RL.