Research On Cloud Computing Task Scheduling Based On Particle Swarm Optimization

With the rapid development of the era,the ever-changing information technology and the rapid advance of the Internet technology,the continuous rise of cloud computing technology,have brought human beings into the "cloud era".Cloud computing attracts the attention of enterprises and academe for its advantages,such as high efficiency,convenience and scalability.Cloud computing builds a pool of resources through virtualization technology,which stores servers,storage and so on,to provide users with on-demand services.Cloud computing task scheduling strategy not only affects the operational efficiency of cloud data center but also affects the satisfaction of the user to the service.Therefore,efficient and reasonable task scheduling strategy is the key to the research of cloud computing.In order to increase satisfaction of cloud user with the efficiency of task scheduling,In this thesis particle swarm algorithm is integrated into cloud computing task scheduling for research and improvement.The main contents are as follows:1.The cloud task scheduling of improved particle swarm optimization and teaching and learning hybrid algorithm is proposed.Aiming at the shortcomings of poor precision and easily falling into local optimum in particle swarm optimization algorithm,put forward the effective information of particle swarm optimization algorithm,which is mixed with the teaching-learning based optimization algorithm with improving teaching factors.The strategy of double sub-populations is applied to the iterative process to share the useful information and improve operating efficiency.And the elimination strategy is introduced.And the improved algorithm is applied to cloud computing task scheduling.The test verification shows that the algorithm scheduling results and the convergence speed is faster in this paper.2.The cloud task scheduling of improved normal updating centroid particle swarm optimization is proposed.First of all,the particle swarm optimization algorithm with normal update mechanism is proposed to avoid the phenomenon of "prematurity".Secondly,the centroid particle is introduced into the algorithm to enhance the particle search ability and increase the particle diversity.Finally,the algorithm can be apply in cloudsim,and the results show that the algorithm is effective.3.By introducing cloning and mutation strategy,the disadvantages of themulti-objective particle swarm optimization algorithm are improved and applied to the cloud.The results show that the improved multi-objective optimization algorithm has more uniform Pareto front-end and have a better scheduling result on the cloud platform.