A Cross-Edge Offloading Framework for Green MEC Systems

This work presents a novel offloading approach to mobile edge computing (MEC). MEC aims to bring computation, storage, and network control closer to the edges of the network, such as base stations and access points, to support resource-limited mobile devices like smartphones and wearable devices. Conventional MEC solutions assume a priori knowledge on the number of processing cycles required by a computational task, which is often impractical. To address this challenge, we propose an online energy-efficient auction (OEA) algorithm for MEC systems consisting of multiple servers and mobile devices. The OEA algorithm takes into account computing tasks with both certain and uncertain processing cycle numbers, and considers both energy consumption and energy harvesting. OEA adopts a combinatorial optimization framework, translated into a series of one-shot procurement auctions optimized using the Vickrey-Clarke-Groves mechanism. The performance of the OEA algorithm is validated through theoretical analysis and simulation studies, demonstrating that it outperforms benchmark strategies.