A Synergistic Hybrid SA-PSO-Double Q-Learning Algorithm for Dynamic Cloud Load Balancing.
Gespeichert in:
| Titel: | A Synergistic Hybrid SA-PSO-Double Q-Learning Algorithm for Dynamic Cloud Load Balancing. |
|---|---|
| Autoren: | N., Ajay, H. S., Mohan |
| Quelle: | International Journal of Intelligent Engineering & Systems; 2025, Vol. 18 Issue 3, p878-893, 16p |
| Schlagwörter: | ARTIFICIAL intelligence, VIRTUAL machine systems, SIMULATED annealing, REINFORCEMENT learning, ALGORITHMS, PARTICLE swarm optimization |
| Abstract: | Efficient task scheduling and load balancing are critical for optimizing resource utilization in cloud computing environments. The current advancement has reported that the optimization and machine learning based artificial intelligence method have gained huge attention in cloud-based application. This article introduces a novel hybrid algorithm that combines Simulated Annealing (SA) with Particle Swarm Optimization (PSO) and Double Qlearning to enhance load balancing and task scheduling in cloud computing environments. In this approach, Simulated Annealing is employed to explore the search space efficiently, leveraging its probabilistic approach to avoid local optima and achieve optimal resource utilization. Particle Swarm Optimization, inspired by the collective behaviour of swarms, refines the search by guiding particles towards the best solutions found so far, ensuring efficient resource optimization. To further enhance decision-making, Double Q-learning, a reinforcement learning technique, is integrated to mitigate overestimation bias by maintaining two Q-value tables, resulting in improved decision accuracy. Additionally, experience replay stabilizes the learning process by breaking temporal correlations between consecutive experiences. The proposed algorithm dynamically schedules tasks across virtual machines (VMs), evaluating parameters such as task execution time, resource utilization, and cost. The performance of proposed approach is validated on synthetically generated data where 100 tasks are generated ranging from 1000 -70k million instructions and GoogleCluster dataset. The obtained performance is measured in terms of average response time, average throughput, energy consumption and the experimental analysis has reported the performance as 5.21s, 96.60 Mbps, and 0.23 kWh, respectively. The result analysis shows that the proposed SAPSO has outperformed existing schemes such as FCFS, SJF, RR, ACO, and PSO algorithms. [ABSTRACT FROM AUTHOR] |
| Copyright of International Journal of Intelligent Engineering & Systems is the property of Intelligent Networks & Systems Society and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Datenbank: | Complementary Index |
Nájsť tento článok vo Web of Science Schreiben Sie den ersten Kommentar!