Research Article
A Novel Deep Federated Learning-Based and Profit-Driven Service Caching Method
@INPROCEEDINGS{10.1007/978-3-031-54531-3_7, author={Zhaobin Ouyang and Yunni Xia and Qinglan Peng and Yin Li and Peng Chen and Xu Wang}, title={A Novel Deep Federated Learning-Based and Profit-Driven Service Caching Method}, proceedings={Collaborative Computing: Networking, Applications and Worksharing. 19th EAI International Conference, CollaborateCom 2023, Corfu Island, Greece, October 4-6, 2023, Proceedings, Part III}, proceedings_a={COLLABORATECOM PART 3}, year={2024}, month={2}, keywords={service caching profit maximization popularity prediction caching decisions collaborative mechanism}, doi={10.1007/978-3-031-54531-3_7} }- Zhaobin Ouyang
Yunni Xia
Qinglan Peng
Yin Li
Peng Chen
Xu Wang
Year: 2024
A Novel Deep Federated Learning-Based and Profit-Driven Service Caching Method
COLLABORATECOM PART 3
Springer
DOI: 10.1007/978-3-031-54531-3_7
Abstract
Service caching is an emerging solution to addressing massive service request in a distributed environment for supporting rapidly growing services and applications. With the explosive increases in global mobile data traffic, service caching over the edge computing architecture, Mobile edge computing (MEC), emerges for alleviating traffic congestion as well as for optimizing the efficiency of task processing. In this manuscript, we propose a novel profit-driven service caching method based on a federated learning model for service prediction and a deep reinforcement learning mode for yielding caching decisions (FPDRD) in an edge environment. The proposed method is temporal service popularity and user preference-aware. It aims to ensure quality of service (QoS) of delivery of cached service while maximizing the profits of network service providers. Experimental results clearly demonstrate that the FPDRD method outperforms traditional methods in multiple aspects.
