Skip to content
SDGs Center UDINUS

Evidence record · UDINUS · SDGs

Novel Unsupervised Cluster Reinforcement Q-Learning for Minimizing Energy Consumption in Federated Edge Cloud

Rektorat UDINUS 2025 Published 2026-10-30

Evidence record · UDINUS · SDGs

Unit
Rektorat UDINUS
Period
2025
Goals
Record
bd3fbc41-2c4c-47bb-bfe2-01ed7ae0a885
Published
2026-10-30

Universitas Dian Nuswantoro conducted research on minimizing energy consumption in Federated Edge Cloud environments through a novel Unsupervised Cluster Reinforcement Q-Learning method. The activity involved UDINUS researchers Shidik, Guruh Fajar; Setiono, Oki; Kusuma, Edi Jaya; Handoko, L. Budi; Andono, Pulung Nurtantio; and Abdollah, Mohd Faizal. The research addressed the challenge of managing distributed energy consumption as global demand for cloud computing increases and green computing becomes essential.

The activity introduced Unsupervised Cluster Reinforcement Q-Learning in Federated Edge Cloud, identified as UCRL-FEC. The method integrates Fuzzy C-Means or K-Means clustering to identify migratable Virtual Machines from overloaded hosts. It also incorporates a modified reward function in Q-Learning to improve energy efficiency and workload balance. Experimental evaluations examined energy consumption, Service Level Agreement Time per Active Host, Performance Degradation due to Migration, and overall Service Level Agreement Violations. A Friedman test was used to confirm statistically significant improvements in energy efficiency, workload distribution, and system stability compared with baseline methods.

The evaluations demonstrated that UCRL-FEC reduced energy consumption by up to 1.07 percent and improved Service Level Agreement Time per Active Host by up to 1.56 percent over the baseline method. Performance Degradation due to Migration was reduced by up to 9.68 percent, while overall Service Level Agreement Violations were reduced by up to 6.06 percent. The findings indicate improved management of active host resources, reduced service disruptions, enhanced service reliability, and optimized resource allocation. The method balances energy efficiency, performance, and scalability through dynamic resource optimization and supports reductions in operational costs and greenhouse gas emissions.

In relation to the stated goals, the activity contributes research on energy efficiency in cloud-edge infrastructure and addresses reductions in greenhouse gas emissions associated with lower energy consumption. Through this research activity, UDINUS frames intelligent Virtual Machine management as a means of supporting sustainable cloud operations.

Reference https://www.scopus.com/pages/publications/105005843247?origin=resultslist