Energy-proficient Task Offloading in Multi-Access Edge Computing using Lyapunov Optimization

Abstract

Multi-access Edge Computing (MAEC) is an emerging concept that integrates cloud computing capabilities within the edge of mobile networks to reduce latency and improve the quality of service for applications. There is a fundamental challenge for Mobile Edge Computing (MEC) in how to transfer computational activities from mobile devices with limited resources to MEC servers efficiently and, at the same time, maximize performance system indicators such as latency, energy consumption, and server load. This paper proposes a new task offloading paradigm for MEC based on Lyapunov optimization theory. The proposed solution has undergone immense testing and has demonstrated high success rates. The framework relies on online offloading and resource allocation strategies for minimizing the average latency of all jobs over time and considering constraints for the stability of task queues and the long-term energy consumption of mobile devices. This scheme has no a priori knowledge of arrival task and channel data for online task execution. Finally, the framework provides provable performance guarantees and bounds the deviation from the best policy. Simulation results show that the new framework is robust and efficiently reduces latency over conventional policies while satisfying stability and energy constraints.