Multi-objective Scheduling Policy for Workflow Applications in Cloud Using Hybrid Particle Search and Rescue Algorithm

Cloud has been developed as a prominent distributed computing model over the last few years because of its wide array of resources and services that are virtualized, scalable, and on demand. In a distributed environment, coordination of workflow applications is an accepted NP-complete problem; hence, it is hard to derive exact solutions. Because of its dynamic and heterogeneous properties, this happens to be even more difficult in cloud environment. The intention of this work is to improve multi-objective optimization of scientific workflow scheduling based on proposed multi-objective hybrid particle search optimization algorithm (MOHPSO) in cloud computing platform and to propose an effective framework for workflow execution. For initial stage, fuzzy Manhattan distance-based clustering is performed to cluster the cloud resources. After that, enhanced chaotic neural network technique is applied to encrypt the task details for security purpose. In this article, the recent search and rescue optimization algorithm (SAR) is hybridized with popular particle swarm optimization algorithm (PSO) to enhance the exploration as well as search ability of optimization algorithm to create best schedules for workflow requests in cloud environment. Moreover, the scientific workflows like Epigenomics, Montage, and Cybershake with varying amount of task sizes are utilized to perform the scheduling process. CloudSim tool is utilized for the simulation of workflow scheduling problem in cloud. Performance enhancement of proposed methodology in terms of load balance, makespan, and cost is validated by comparison with various state-of-the-art algorithms.

.