Task Scheduling In Cloud Computing Environment Based On NPSO Algorithm

Cloud computing is an emerging technology,it shares infrastructuretechnology based the network, the resource pool is built by a large number of cheapresources node, the performance of each resource node has larger differences, how touse limited resources to provide satisfactory services to users is an important issue,cloud computing need to solve it. Therefore, the task scheduling problem hasbecome a hot research topic in cloud computing.Currently, most cloud computing environments are based on Hadoop, usemaster-slave mode to complete the task scheduling process, and choose the taskscheduling algorithm of Hadoop architecture to be the task scheduling algorithm ofthe cloud computing environments mostly. However, the cloud computingenvironment is a multi-user environment,and it is business-oriented, Hadoopframework scheduling models and algorithms does not fully meet the requirementsof cloud computing.In order to make a better combination between the Hadoop frameworkscheduling models and the specific circumstances of the cloud environment, thepaper improved the master-slave mode of Hadoop, so that the task can becomplicated through the free choice of resource nodes. The Master node and theAssist node work together to complete the task scheduling process.In this paper, the improvements of the standard particle swarm algorithminclude two aspects: first, the idea of "split" is introduced into the algorithm; second,use the asynchronous iterative mechanism to complete the speed and position update.In addition, the paper not only map tasks and resources but only give fullconsideration to the load balancing of cloud resource pool. So,when use particleswarm scheduling algorithm, the paper use objective function and fitness function atthe same time to coordinate the completion of the task scheduling process.Finally, the paper uses the simulation software CloudSim to do the experimentalsimulation,it proves that the using NPSO algorithm to solve the task scheduling problem in the cloud environment can effectively improve the efficiency of taskscheduling and ensure load balancing of cloud resources, while avoiding the"premature" problem of the standard PSO algorithm.