Sustainable Energy Management in Mobile Edge Computing Network with Multi-Agent Actor Critic Algorithm

Recognizing the paramount importance of sustainability in contemporary technological landscapes, the research addresses the challenges posed by mobile edge computing (MEC) applications. The study delves into enhancing the efficiency of energy utilization in densely deployed MEC hosts across wireless networks. To mitigate the energy consumption associated with high-capacity MEC hosts, the strategy involves the integration of self-powered base station (BS) units. The present study introduces an approach through a decentralized energy management system that formulates a dynamic optimization problem to minimize overall energy consumption while meeting the system's energy demand. The proposed method utilizes a multi-agent actor critic algorithm to find the best resource allocation for the network and utilize renewable energy sources. In this approach, each BS acts as an independent agent, making decisions based on its local observations. The collaborative interaction among these agents facilitated an optimized energy dispatch strategy. This strategy is then assessed on various metrics, including energy consumption, energy costs, and overall system performance.