Scheduling Of Cloud Computing Resources Based On Improved Particle Swarm Optimization

Cloud computing is a business model used virtualization and converged network technology to provide users with the service. The rapidexpansion of virtualization technology and universal application of Internet technology make cloud computing become a hot research rapidly. With the expanding of cloud computing resources, the increasing number of cloud users, the increasing difficulties of resource scheduling, so the cloud users’ requirement of Qo S(Quality of Service) is difficult to satisfied, and the cloud load balancing is also very difficult to guarantee, which makes the resource scheduling in cloud face serious challenges. As a consequence, this article focuses on cloud computing resource scheduling algorithms and scheduling model, giving full consideration of users’ needs and load balancing in cloud environments.Detail research content is as follows:(1) Propose ways to improve particle swarm optimization. In view of PSO shortcomings, out of some bad initialization PSO particles dynamically adjust some parameters to develop algorithms to determine the rules and precocious approach, using the penalty function for improved algorithm optimization, analyze the improved particle swarm optimization by matlab.(2) Establish a physical model of cloud computing resource scheduling. Combine the improved particle swarm optimization with cloud computing resources scheduling, establish a physical model physical model to realize cloud computing resource scheduling.This model considers Qo S requirements and tries best effort to meet the needs of cloud users and takes full account of resource scheduling physical machine load, the load factor by limiting physical machine to place the physical resources to keep load balancing in cloud environment.(3) Design experiments on improved particle swarm algorithm applied to cloud computing to analyze the performance of resource scheduling. Cloud Sim simulation experiments using improved particle swarm algorithm to eliminate the availability of resources, resource search, the cloud load balancing situation, the task execution time span in comparison with other resource scheduling algorithm experimentally derived data analysis.