Dynamic Task Scheduling for Heterogeneous Multicore Baseband Processors in IIoT
IEEE Internet of Things Journal · 2026-11 · 13 (22) · 51931-51946
DOI: 10.1109/jiot.2026.3727005 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Internet of Things Journal, 2026-11. Verfasst von Wei Chen und 2 weiteren Autor:innen. Die Arbeit nennt 27 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Natural Science Foundation of Hainan Province, Foreign Expert Project of Hainan Province.
Abstract (Originalsprache)
The scheduling of tasks presents significant challenges in heterogeneous multicore baseband processors for industrial Internet of Things (IIoT) applications: dynamic heterogeneity, task diversity, and multiobjective optimization. Traditional static scheduling methods cannot adapt to fluctuating workloads, while existing deep reinforcement learning (DRL) approaches provide limited support for dynamic, heterogeneous, and personalized scheduling. To overcome these limitations, we propose the heterogeneity-sensitive multimodal policy network (HS-MPN) and a game-theory-driven personalized federated learning framework. On this basis, we develop a dependency-aware dynamic task-scheduling method within an end-to-end scheduling architecture. First, HS-MPN fuses task features, precedence states, and core states through a bidirectional attention mechanism. This mechanism models dynamic task–core interactions to support dependency-aware priority adjustment and load-balanced scheduling. Second, we design a three-layer personalized federated aggregation architecture grounded in game theory. The game-theoretic coordination mechanism balances local adaptation and global collaboration, promoting stable global aggregation while preserving node-specific heterogeneity. Experiments show that, for 1000 tasks, the proposed method reduces task-completion time by 37.64%, energy consumption by 29.17%, and core-load variance by 34.1% relative to the best-performing baseline, while achieving an 86.75% quality-of-service compliance rate. Our method exhibits robustness in high-concurrence communication-algorithm scenarios by deeply coupling dynamic scheduling with federated collaboration. The proposed method provides an effective dynamic task scheduling solution for heterogeneous multicore baseband processors in IIoT.
Förderung
- National Natural Science Foundation of China
- Natural Science Foundation of Hainan Province
- Foreign Expert Project of Hainan Province
- National Foreign Expert Project of China
Zitieren
APA
Chen, W., Peng, J., & Liu, D. (2026). Dynamic Task Scheduling for Heterogeneous Multicore Baseband Processors in IIoT. IEEE Internet of Things Journal, 13(22), 51931-51946. https://doi.org/10.1109/jiot.2026.3727005
Vancouver
Chen W, Peng J, Liu D. Dynamic Task Scheduling for Heterogeneous Multicore Baseband Processors in IIoT. IEEE Internet of Things Journal. 2026;13(22):51931-51946. doi:10.1109/jiot.2026.3727005
BibTeX
@article{Chen2026,
author = {Chen, Wei and Peng, Jie and Liu, Dake},
title = {Dynamic Task Scheduling for Heterogeneous Multicore Baseband Processors in IIoT},
journal = {IEEE Internet of Things Journal},
year = {2026},
volume = {13},
number = {22},
pages = {51931--51946},
doi = {10.1109/jiot.2026.3727005}
}
EUVIMED