一种自适应粒子群算法在云资源调度中的应用

Application of an Adaptive Particle Swarm Algorithm in Cloud Scheduling

摘要:

通过资源调度优化提升云计算的效率并降低数据中心能耗是云计算领域的主要研究内容之一。粒子群算法常用于解决资源调度问题，然而粒子群算法在云计算资源调度应用中算法初期收敛速度快，后期收敛速度缓慢，易陷入局部寻优。本文提出了一种自适应改进的粒子群算法用于云计算资源调度问题的研究，该算法通过自适应改进粒子的个体学习因子和社会学习因子，以提高算法的全局探索能力，使得粒子逼近更优解。实验结果表明：本文提出的自适应粒子群算法不仅具备良好的收敛性和全局寻优能力，同时能够大幅度降低云资源调度中任务队列的总完成时间。

Abstract:

Improving the efficiency of cloud computing and reducing the energy consumption of data center is one of the main research contents in cloud computing. Particle swarm optimization (PSO) is often used to solve resource scheduling problems. However, in the application of cloud computing resource scheduling, PSO has fast initial convergence speed, slow convergence speed and easy to fall into local optimization. in this paper, we propose an adaptive improved particle swarm optimization algorithm for cloud computing resource scheduling problem. the algorithm improves the individual learning factors and social learning factors of particles by adaptive improvement, in order to improve the global exploration ability of the algorithm and make the particles approximate the better solution. The experimental results show that the proposed adaptive PSO not only has good convergence and global optimization ability, but also can greatly reduce the total completion time of task queue in cloud resource scheduling.