A Comparative Study of Swarm Intelligence (SI) Task Scheduling Algorithms in Cloud Computing

Abstract

Cloud Computing refers to computing services over the internet and deals with varied different virtualization resources. The task scheduling plays a crucial role in enhancing the performance of cloud computing. The issue with task scheduling is distribution of tasks within the system in a manner that optimize the performance of overall system and minimize the execution time. To achieve such good plan the provider need to evaluate and choose among different algorithms to allocate and schedule the available resources. The challenging decision of choosing the proper algorithm is taking based on different performance metrics for task scheduling. In this research the focus is concentrated on task Execution Time as criteria for evaluating among the chosen algorithms. The selected mechanisms contain information of jobs (cloudlets) and resources (virtual machines) such as length of jobs, speed of resources and identifier for both. In order to generate the population, first, set of jobs and resources were created, then the execution times of jobs were computed as a fitness values. Second, the algorithms iterated to themselves in order to regenerate populations to produce the best job schedule that gives the minimum execution time of jobs. The methodology of this research is based on simulation of the selected mechanisms using the Java Language and CloudSim simulator. The comparison and analysis of different task scheduling algorithms has been discussed in this research on the basis of time execution. The results revealed that when having small sizes of scheduling problems PSO take the lead. However, in case of large size of jobs, Cat Swarm Optimization significantly outperforms the considered Particle Swarm Optimization, Firefly Algorithm and Glowworm Swarm Optimization in terms of execution time.