A comparative study into swarm intelligence algorithms for dynamic tasks scheduling in cloud computing

Cloud computing are becoming the major source of computing. The core idea of this environment is managing and scheduling the available resources to provide service's needs. Servers in cloud may be physical or virtual machines accessed across the network. Selecting machines for executing a task in the cloud computing must be considered. They have to be selected according to its status and submitted tasks properties to exploit the efficiency of the resources. Cloud task scheduling is considered an NP-hard optimization problem, and many meta-heuristic algorithms are suitable to solve it. This paper investigates three possible approaches proposed for dynamic task scheduling in cloud computing. The three approaches are belonging to the field of swarm intelligence that is used to find solutions for difficult or impossible combinatorial problems. These approaches are inspired by ant colony behavior, the behavior of particle swarm and honeybee foraging behavior. The main goal is to provide an evaluation and comparative study of these approaches that are used to minimize the makespan of a given tasks set. Performance of the algorithms is simulated using toolkit package of CloudSim. Algorithms have been compared with each other and with the well-known existed algorithms for dynamic task scheduling problem. The results of the experiments are presented and the strengths of each algorithm are investigated. Experimental results show that the proposed approaches satisfy expectation, also proved that ABC algorithm is the superior than other algorithms.