Accepted author manuscript, 3.71 MB, PDF document
Available under license: CC BY: Creative Commons Attribution 4.0 International License
Final published version
Research output: Contribution to Journal/Magazine › Journal article › peer-review
Research output: Contribution to Journal/Magazine › Journal article › peer-review
}
TY - JOUR
T1 - Edge Data Caching With Consumer-Centric Service Prediction in Resilient Industry 5.0
AU - Liu, Guoqiang
AU - Bao, Guanming
AU - Bilal, Muhammad
AU - Jones, Angel
AU - Jing, Zhipeng
AU - Xu, Xiaolong
PY - 2024/2/29
Y1 - 2024/2/29
N2 - Industry 5.0, emerging as a promising industry paradigm, unleashes the potential of improving consumer experience by delivering consumer-centric services, facilitating substantial growth in consumer electronics. To improve the resilience of industry 5.0, edge data caching enables sustainable and low-latency service provision by caching data at edge servers (ESs) closer to production. However, the limited caching capacity of ESs presents a formidable challenge to efficient edge data caching. Moreover, the dynamic of consumer-centric service requests further complicates the effective implementation of caching strategies. In response to the above challenges, we propose an edge data caching scheme, named SPM-ECDP, with consumer-centric service prediction for Industry 5.0. Initially, a time-series prediction model is employed to forecast the service demands. To ensure the confidentiality of data, federated learning is introduced in the model training phase. Subsequently, reinforcement learning is adopted to enable ESs to make intelligent decisions on edge data caching, consequently enhancing caching efficiency. Through comprehensive simulation experiments, the effectiveness and superiority of the proposed scheme in increasing caching hit ratio and reducing data delivery delays are demonstrated. The experimental results demonstrate that the proposed SPM-ECDP method has enhanced the hit ratio by 7.05% -48.5% when compared to the baseline method.
AB - Industry 5.0, emerging as a promising industry paradigm, unleashes the potential of improving consumer experience by delivering consumer-centric services, facilitating substantial growth in consumer electronics. To improve the resilience of industry 5.0, edge data caching enables sustainable and low-latency service provision by caching data at edge servers (ESs) closer to production. However, the limited caching capacity of ESs presents a formidable challenge to efficient edge data caching. Moreover, the dynamic of consumer-centric service requests further complicates the effective implementation of caching strategies. In response to the above challenges, we propose an edge data caching scheme, named SPM-ECDP, with consumer-centric service prediction for Industry 5.0. Initially, a time-series prediction model is employed to forecast the service demands. To ensure the confidentiality of data, federated learning is introduced in the model training phase. Subsequently, reinforcement learning is adopted to enable ESs to make intelligent decisions on edge data caching, consequently enhancing caching efficiency. Through comprehensive simulation experiments, the effectiveness and superiority of the proposed scheme in increasing caching hit ratio and reducing data delivery delays are demonstrated. The experimental results demonstrate that the proposed SPM-ECDP method has enhanced the hit ratio by 7.05% -48.5% when compared to the baseline method.
KW - Consumer-centric
KW - Data models
KW - Data privacy
KW - Federated learning
KW - Industries
KW - Production facilities
KW - Servers
KW - Training
KW - edge caching
KW - edge computing
KW - industry 5.0
KW - privacy preservation
KW - reinforcement learning
U2 - 10.1109/tce.2023.3327847
DO - 10.1109/tce.2023.3327847
M3 - Journal article
VL - 70
SP - 1482
EP - 1492
JO - IEEE Transactions on Consumer Electronics
JF - IEEE Transactions on Consumer Electronics
SN - 0098-3063
IS - 1
ER -