Computation Task Scheduling for Computing Energy Efficiency in Heterogeneous Edge Computing Systems

With the exponential increase in demand for computing power across cloud, edge, and device in modern society, mobile edge computing (MEC) has garnered renewed attention from both academia and industry, attributed to its rapid response capabilities. However, as application scenarios diversify and the generation gap in edge computing infrastructure widens, the complexity of computing task scheduling has significantly increased. Meanwhile, the integration of distributed energy with heterogeneous edge computing resources imposes higher demands on computational task scheduling. The synergistic optimization of power supply, task scheduling, and edge networks has emerged as a key research focus for the future. In this paper, we propose a framework for an edge computing system that integrates distributed energy resources with heterogeneous computing capabilities at the network edge. Subsequently, we formulate the energy efficiency of the edge computing system as the optimization objective and propose a novel computing task scheduling method by leveraging deep reinforcement learning. Finally, the simulation results demonstrate that the proposed method achieves efficient integration of edge computing systems with distributed energy resources through optimized task scheduling, thereby significantly enhancing the system’s energy efficiency.