云环境下基于节能和负载均衡的混沌粒子群资源优化调度

Chaos Particle Swarm Optimizing Scheduling Based on Power-aware and Load Balance in Cloud Computing

摘要:

针对传统云计算资源调度方法仅关注任务的最大完成时间,没有考虑到节能和资源负载均衡的问题,提出了一种基于混沌粒子群算法实现云资源优化调度的方法；首先,定义了以节能和负载均衡为目标的多目标数学模型,然后设计了一组靠近最优Pareto 前沿的解作为初始种群,采用改进的粒子群算法来搜索最优调度方案,当最优解连续两代未发生变化时,通过混沌遍历法对粒子进行局部寻优,以加快获取全局最优解；在CloudSim仿真环境下结合Matlab工具进行实验,结果表明:文中方法负载均衡离差平均值为0.156,且较其它方法,具有较好的负载均衡能力和较低的能耗,具有很强的可行性。

Abstract:

Aiming at the traditional cloud computing resource scheduling method only considering the latest finishing time of task, not concerning the power-aware and resource load balance, a method based on chaos particle swarm was proposed. Firstly, the mathematical model based on energy-saving and load balance was built, then a set of solution near Pareto front as the initial population, the improved particle swarm algorism was used to search the optimal scheduling scheme. When the optimal solution was not changed for two iterations, the chaos search was used to search to get the global optimum solution. The simulation in the CloudSim and Matlab environment shows the method in this paper has the average load balance distance as 0.156, and compared with the other methods, it has the good load balance performance and low energy consumption. It has strong feasibility.