RAA-DRL: Renewable-Aware AI-Driven Resource Allocation for Green Communications and Energy-Efficient Wireless Networks

Abstract

The continual rise in the demand for greater capacity from such systems has evoked concern regarding energy consumption and ultimately, an environmental impact. Traditional resource allocation methods generally prioritize performance and overlook the unpredictability of renewable energy; therefore, such methods could lead to waste and the use of non-renewable resources. To help tackle these aspects, this paper offers RAA-DRL (Renewable-Aware AI-Driven Resource Allocation), an innovative framework for optimizing network performance while incorporating green energy. RAA-DRL utilizes deep reinforcement learning (DRL) to allocate resources in real-time, based on the current traffic and available level of renewable energy. Unlike static methods, the DRL-based system will learn dynamically from historical traffic and renewable generation data regarding long-term energy efficiency, throughput and spectral efficiency. The simulation results show that RAA-DRL can deliver significant improvements in energy efficiency and QoS compared to conventional schemes. It can reduce the reliance on the grid and continue to support high network performance, serving as an example of what AI technology can offer for sustainable communications. This research contributes to the establishment of green communication technologies through a scalable, adaptive, and renewable-aware approach to future wireless networks. Ultimately, RAA-DRL enables viable, sustainable, performance oriented networks for the future 6G and IoT systems.