Modified parallel PSO algorithm in cloud computing for performance improvement

For accessing required services in cloud computing, user submits its task to the cloud datacentre for processing. Therefore, two challenges have been faced by datacentre controllers such as finding the best resources and mapping user tasks to virtual machines (VMs). To solve these issues, this paper presented a scheduling algorithm named as Modified Parallel Particle Swarm Optimization (MPPSO). This algorithm is based on the Parallel PSO algorithm which reduces the processing time and dynamically adjust the load of each VM that VM can take part in task processing. By using the CloudSim simulator, MPPSO approach is tested against Parallel Particle Swarm Optimization (PPSO) and Modified Particle Swarm Optimization (MPSO) algorithm by taking different task and VM sets. From the result our proposed algorithm reduce execution time, makespan time and waiting time by 16%, 15% and 19% while increase the throughput and fitness function value by 16% and 17% respectively.