A novel multi-objective CR-PSO task scheduling algorithm with deadline constraint in cloud computing

• Published in: Sustainable computing informatics and systems Vol. 32; p. 100605
• Main Authors: Dubey, Kalka, Sharma, S.C.
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01.12.2021
• Subjects: Cloud computing; CRO; Deadline; Multi-objective; PSO; Task scheduling; Task scheduling; Cloud computing; CRO; Multi-objective; PSO; Deadline
• ISSN: 2210-5379
• DOI: 10.1016/j.suscom.2021.100605

•A multi-objective hybrid method for task scheduling problem is proposed.•Represented the NP-Complete task scheduling problem in the form of mathematical model.•Proposed a task scheduling framework for processing the applications/tasks in efficient way.•The proposed method is a hybrid optimization technique using the Chemical Reaction Optimization (CRO) and Particle Swarm Optimization (PSO).•Proposed CR-PSO framework is simulated at CloudSim simulator to evaluated the performance. In cloud computing, efficient task scheduling espouses many challenges. To schedule the multiple cloudlets with deadline constraints on hybrid cloud resources while meeting the various quality requirements is a challenging issue. The purpose of this research work is to address the task scheduling problem of cloud computing. A novel hybrid task scheduling algorithm named Chemical Reaction Partial Swarm Optimization has been proposed for the allotment of multiple independent tasks on the available virtual machines. It enhances the classical chemical reaction optimization and partial swarm optimization and does hybridization by combining the features for the optimal schedule sequence where tasks can be processed based upon the demand and deadline simultaneously to improve the quality in terms of factors like cost, energy, and makespan. We present the comprehensive simulation experiment using the CloudSim toolkit, which shows the effectiveness of the proposed algorithms. To analyse average execution time, comparative experiments have been carried out using various combinations of virtual machines and the number of tasks. The results bring out a significant reduction in execution time of the order of 1–6 percent, which further improves even more than 10 percent in some cases. The results of the makespan reflect the effectiveness of the algorithm in order of 5–12 percent, and the outcome of total cost 2–10 percent and energy consumption rate shows the 1–9 percent improvement.