• Ravi Prasad Ravuri

In this article, we proposed a learning based methodology for task scheduling in cloud computing towards leveraging performance of cloud infrastructure. It also enables consumers to have their Service Level Agreements (SLAs) satisfied with resource optimization. Deep Reinforcement Learning (DRL) technique is used in the proposed methodology for making scheduling decisions. Towards this end an algorithm known as Intelligent Learning based Task Scheduling (ILTS)is proposed and implemented. The proposed algorithm exploits agent based iterative approach that makes use of action-space, state-space and reward function to make well informed scheduling decisions. With the action-feedback iterations, the algorithm can make accurate scheduling decisions that improve energy efficiency and improve Quality of Service (QoS) in execution of jobs in cloud. DRL involves state transition in each stage and there is update of its Q-table through underlying deep neural network. We made experiments with different workloads. Our empirical study has revealed that the proposed system is better than existing methods in terms of success rate, energy efficiency and execution time.

Task Scheduling, Service Level Agreements, Cloud Computing, Machine Learning, Deep Reinforcement Learning

"A Deep Reinforcement Learning Based Framework for Task Scheduling for Enhancing Efficiency in Cloud Computing", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.10, Issue 6, page no.i643-i651, June-2023, Available :http://www.jetir.org/papers/JETIR2306869.pdf

"A Deep Reinforcement Learning Based Framework for Task Scheduling for Enhancing Efficiency in Cloud Computing", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.10, Issue 6, page no. ppi643-i651, June-2023, Available at : http://www.jetir.org/papers/JETIR2306869.pdf