AI-Enabled Framework for Energy Sustainability Evaluation and Enhancement of Wireless Networks

In the pursuit of sustainable wireless connectivity, this paper leverages artificial intelligence (AI) to evaluate and enhance the energy sustainability of wireless networks. To begin, energy sustainability is astutely defined, encompassing both the short-term metric and, more critically, long-term evaluations, all while taking careful consideration of the dynamics of the wireless operation environment. Then, treating the task at hand as a sequential Markov decision process, an AI-driven framework powered by quantum reinforcement learning (QRL) is proposed to evaluate the long-term energy sustainability and leverage it to maximize the short-term sustainability, particularly in terms of the instantaneous energy efficiency. A case study is also presented, where the framework is deployed within the context of an aerial base station aided wireless communication network, showcasing the effectiveness and real-world applicability of the proposed framework. The results demonstrate significant reductions in energy consumption and improvements in energy efficiency versus the approach devoid of long-term prediction, underscoring the notable advantages of the new framework. Furthermore, the QRL-based solution consistently outperforms conventional deep reinforcement learning techniques, highlighting its superior capability in balancing short-term performance with long-term sustainability objectives. It also demonstrates advantages in terms of faster convergence in dynamic wireless environments and efficient operation in high-dimensional state and action spaces.