TY - GEN
T1 - A Multi-agent Reinforcement Learning based Offloading Strategy for Multi-access Edge Computing
AU - Ma, Li
AU - Shi, Haobin
AU - Li, Jingchen
AU - Hwang, Kao Shing
N1 - Publisher Copyright:
© 2021 IEEE
PY - 2021
Y1 - 2021
N2 - As a front-end distributed computing paradigm, Multi-access edge computing (MEC) requires an efficient offloading strategy in the case of computation-intensive tasks. Traditional heuristic and reinforcement learning-based methods are limited by the number of edge serves. In this work, the task offloading in MEC is regarded as a multi-agent reinforcement learning (MARL) scenario, and an end-to-end model is developed to train the offloading strategy. The proposal is a decentralized framework, which is consistent with the edge computing servers. We fully consider the state sequences for edge servers and limited communication abilities, proposing a recurrent neural network as the communication module. Using a gating mechanism, we design a dual-recurrent network for combining the state sequences and historical communication results. Several experiments show that our method can achieve low latency and distributed data processing, and it outperforms heuristic methods and other reinforcement learning-based frameworks.
AB - As a front-end distributed computing paradigm, Multi-access edge computing (MEC) requires an efficient offloading strategy in the case of computation-intensive tasks. Traditional heuristic and reinforcement learning-based methods are limited by the number of edge serves. In this work, the task offloading in MEC is regarded as a multi-agent reinforcement learning (MARL) scenario, and an end-to-end model is developed to train the offloading strategy. The proposal is a decentralized framework, which is consistent with the edge computing servers. We fully consider the state sequences for edge servers and limited communication abilities, proposing a recurrent neural network as the communication module. Using a gating mechanism, we design a dual-recurrent network for combining the state sequences and historical communication results. Several experiments show that our method can achieve low latency and distributed data processing, and it outperforms heuristic methods and other reinforcement learning-based frameworks.
KW - Edge computing
KW - Multi-agent reinforcement learning
KW - Reinforcement learning
UR - http://www.scopus.com/inward/record.url?scp=85123941357&partnerID=8YFLogxK
U2 - 10.1109/CACS52606.2021.9639048
DO - 10.1109/CACS52606.2021.9639048
M3 - 会议稿件
AN - SCOPUS:85123941357
T3 - 2021 International Automatic Control Conference, CACS 2021
BT - 2021 International Automatic Control Conference, CACS 2021
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2021 International Automatic Control Conference, CACS 2021
Y2 - 3 November 2021 through 6 November 2021
ER -