Abstract:

Load balancing technology is always the important part of cloud resource management, load balancing should be used in management and maintenance of the data center, which can improve the efficiency of resource use, reduce the number of virtual machine migration effectively, avoid the system bottleneck. Most of the existing virtual machine deployment algorithms based on particle swarm integrate the multiple performance objective of the virtual machine into one target by weight assignment when calculating the fitness. Moreover, in the process of updating the position of particles, the method to approach the current optimal solution is very easy to make the final solution trap into the local optimum. In view of the above problems, a hybrid particle swarm optimization based on multi-objective non dominated solution is proposed to solve the load balancing problem of virtual machines. The simulation results show that the proposed algorithm is more effective than the general particle swarm algorithm in load balancing.

Key words: virtual machine deployment, hybrid particle swarm optimization, load balancing