Multi-Layer Computation Offloading in Distributed Heterogeneous Mobile Edge Computing Networks
- 24 March 2022
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive Communications and Networking
- Vol. 8 (2) , 1301-1315
Abstract
In this paper, we consider distributed heterogeneous multi-layer mobile edge computing (HetMEC) networks, where resource-poor edge devices (EDs) upload computing tasks for processing to the mobile edge computing (MEC) servers and a cloud center (CC). To reduce total energy consumption, computation offloading and resource allocation are independently performed by each device and each server. However, due to the partial information available at each device and server, the offloading strategies may overwhelm the layers above. This may lead to network congestion, i.e., so many tasks are offloaded to the same node that this node is overloaded. To address this problem, we develop a smart pricing mechanism to coordinate the computation offloading of multi-layer devices, where the CC charges the MEC servers and EDs for computing services and network congestion. In particular, to satisfy the latency constraints of each task, we construct a Lagrangian framework where multi-agent reinforcement learning is utilized by each MEC server to determine its offloading strategies and resource allocation, so that the total energy consumption is reduced. Simulation results show that our algorithm achieves an energy consumption reduction of 28% and a decrease in congestion probability of between 28% and 100% compared to the state of the art.Keywords
Funding Information
- National Nature Science Foundation of China (61941101, 61931019)
- Beijing Natural Science Foundation (L212027)
This publication has 31 references indexed in Scilit:
- Mobile edge computing towards 5G: Vision, recent progress, and open challengesChina Communications, 2016
- Smart data pricingCommunications of the ACM, 2015
- Game-theoretic Analysis of Computation Offloading for Cloudlet-based Mobile Cloud Computing2015
- Efficient Multi-User Computation Offloading for Mobile-Edge Cloud ComputingIEEE/ACM Transactions on Networking, 2015
- Joint Optimization of Radio and Computational Resources for Multicell Mobile-Edge ComputingIEEE Transactions on Signal and Information Processing over Networks, 2015
- Evolution toward 5G multi-tier cellular wireless networks: An interference management perspectiveIEEE Wireless Communications, 2014
- Cooperative Multi-Agent Learning: The State of the ArtAutonomous Agents and Multi-Agent Systems, 2005
- Rayleigh fading channels in mobile digital communication systems .I. CharacterizationIEEE Communications Magazine, 1997
- Processor design for portable systemsJournal of Signal Processing Systems, 1996
- Augmented Lagrange Multiplier Functions and Duality in Nonconvex ProgrammingSIAM Journal on Control, 1974