An Improved Optimal and Dynamic Resource Allocation Model for Cloud Services Using Machine Learning

Hari Krishnan Andi; Divya; Hishamuddin Bin M.Salleh

doi:10.69980/redvet.v25i1.2145

Hari Krishnan Andi
Divya
Hishamuddin Bin M.Salleh

DOI: https://doi.org/10.69980/redvet.v25i1.2145

Keywords: Cloud Computing, Resource Management, Workload Prediction, Task Scheduling, Virtual Machine Allocation, Machine Learning Optimization

Abstract

Cloud computing enables elastic and on-demand provisioning of computational and storage resources. However, dynamic workloads pose challenges in task scheduling, VM allocation, and resource management. Inefficient allocation may lead to underutilization, increased energy consumption, and SLA violations. This study introduces a unified framework integrating Pulse-Coupled Genetic Particle Swarm Optimization Neural Network (PCGPSONN) with Support Vector Machine (SVM) for resource demand prediction, Phasmatodea Population Modified McNaughton Evolution (PPMMcNE) for task scheduling, and Rat Swarm Modified Brucker Optimization (RSMBO) for VM allocation. Experimental results demonstrate forecasting accuracy of 96.29%, throughput of 0.942, and significant reductions in execution cost and energy usage. This paper contributes a comprehensive machine learning–optimization approach to sustainable cloud resource management.

Author Biographies

Hari Krishnan Andi

Faculty of Computer Science and Multimedia, Lincoln University College, Malaysia

Divya

Faculty of Computer Science and Multimedia, Lincoln University College, Malaysia,

Hishamuddin Bin M.Salleh

Faculty of Computer Science and Multimedia, Lincoln University College, Malaysia ,

References

1. Anupama, R., Shivakumar, R., & Nagaraja, G. (2021). Hybrid models for resource utilization prediction in cloud computing. *International Journal of Cloud Applications*, 12(3), 45–59.
2. Kumar, A., Sahoo, P., & Gupta, R. (2021). Self-directed learning-based workload prediction for efficient cloud management. *IEEE Transactions on Cloud Computing*, 9(4), 556–567.
3. Khan, A., Li, H., & Zhang, P. (2022). Machine learning centric resource management in cloud data centers. *Future Generation Computer Systems*, 127, 1–13.
4. Behera, S., & Sobhanayak, S. (2024). Hybrid GA-GWO task scheduling optimization for heterogeneous cloud systems. *Journal of Parallel and Distributed Computing*, 180, 45–57.
5. Estrada, F., Kim, J., & Patel, V. (2023). CPU usage prediction with VM clustering for efficient resource management. *ACM Transactions on Cloud Computing*, 11(2), 120–135.
6. Geetha, R., Bhatia, A., & Rao, K. (2024). Hybrid optimization algorithms for efficient workload distribution in cloud systems. *IEEE Access*, 12, 8851–8865.
7. Gupta, H., Verma, P., & Singh, K. (2021). Hybrid heuristic approaches for cloud scheduling problems. *Concurrency and Computation: Practice and Experience*, 33(18), e5671.
8. Khurana, M., Sharma, A., & Das, S. (2023). Hyperparameter-tuned gradient boosting for CPU utilization prediction. *International Journal of Big Data Analytics*, 10(1), 67–82.
9. Malik, T., Singh, A., & Reddy, P. (2022). Evolutionary algorithm and machine learning for cloud data center utilization. *Journal of Systems Architecture*, 124, 102381.
10. Mishra, S., Kumar, V., & Das, A. (2020). Energy-aware task allocation in multi-cloud environments. *Journal of Cloud Computing*, 9(14), 112–126.
11. Mohan, R., & Kangasharju, J. (2021). Cloud workload prediction for efficient resource allocation. *Future Internet*, 13(9), 232.
12. Patel, N., Swain, R., & Chen, J. (2023). Cloud resource management: Challenges and emerging solutions. *ACM Computing Surveys*, 55(6), 115–139.
13. Sharma, P., Kumar, S., & Gupta, M. (2020). Machine learning for dynamic cloud resource allocation. *International Journal of Cloud Engineering*, 8(2), 101–119.
14. Swain, R., Nayak, R., & Patnaik, A. (2022). Optimal VM allocation strategies in cloud computing. *IEEE Transactions on Services Computing*, 15(3), 1441–1452.
15. Zhang, Y., Yao, J., & Guan, X. (2017). Deep reinforcement learning for cloud resource management. *IEEE Transactions on Cloud Computing*, 5(4), 762–774.