Adaptive scheduling algorithms for energy efficiency in hybrid cloud-edge computing paradigms
1 College of Technology, University of North America (UONA).
2 Computer communication and information sciences, Aalto University School of Science, Aalto University.
Research Article
International Journal of Science and Technology Research Archive, 2026, 10(01), 073-085.
Article DOI: 10.53771/ijstra.2026.10.1.0067
Publication history:
Received on 13 November 2025; revised on 13 March 2026; accepted on 28 March 2026
Abstract:
Hybrid cloud-edge computing faces energy-efficiency challenges, as workloads can be dynamic and resource provisioning may be heterogeneous, leading to varying resource utilization. This research introduces ADEES, an Adaptive Dynamic Energy-Efficient Scheduler, that dynamically minimizes energy consumption while meeting QoS requirements in distributed environments. ADEES does this by combining real-time workload classification with resource allocation through reinforcement learning to optimize task distribution between edge and cloud nodes. The results indicate that ADEES outperforms conventional schedulers (i.e., Round-Robin, Kubernetes Default, and genetic algorithm) in simulations based on replicated Alibaba cluster traces and synthetic workloads, achieving 22.4% lower energy consumption and 9.8% more completed tasks, while incurring only a 7% increase in latency for non-critical tasks. ADEES is up to 2000 times lighter than other approaches and scales efficiently for deployments of up to 500 nodes. This research contributes to sustainable distributed computing by investigating the critical trade-off between energy efficiency and application performance in the context of 5G and the Internet of Things. The key contributions are the two-tiered adaptation strategy and the practical recommendations for heterogeneous environments.
Keywords:
Energy-efficient scheduling; Cloud-edge computing; Reinforcement learning; Resource allocation and distributed systems
Full text article in PDF:
Copyright information:
Copyright © 2026 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0