RAQPSO算法的云计算资源调度策略

在"互联网+"时代,云计算代表了一种新的商业模式,而云系统中用户任务与计算节点的调度问题极大地影响着系统的性能和云竞争力。为此,提出了一种改进的量子粒子群算法——反向自适应量子粒子群算法(RAQPSO),通过对惯性权值参数的调整和加入反向学习算子来提高算法的全局搜索能力,并将其应用于云计算资源调度中,仿真验证了算法的有效性。建立了云计算资源调度问题的模型;采用自适应机制,将适应度函数的变化程度作为惯性权值的更新因子,避免了单纯地根据迭代次数的线性函数来取值,从而使粒子不易陷入局部最优;随后加入粒子反向学习算子,加强了粒子全局搜索能力。实验结果表明,RAQPSO算法大大节约了任务完成时间,并且保持了良好的计算节点负载平衡。

Cloud Computing Resource Scheduling Strategy Based on Improved RAQPSO Algorithm Research

In ''INTERNET + TIME'', cloud computing represents a novel business model. However, the cloud user tasks in the system and compute node scheduling problem significantly affects the system performance and competitiveness of cloud. An improved algorithm of quantum particles adaptive quantum particle swarm optimization (RAQPSO), based on the inertia weight adjustment of parameters and reverse learning to improve the global search ability of the algorithm, and applied to cloud computing resource scheduling problem to verify the effectiveness of the algorithm. With cloud computing resource scheduling model is established. And then uses the adaptive mechanism, the change of the fitness function as update of inertia weight factor, avoids simply value according to the linear function of the number of iterations. Add the particle reverse learning operator, to strengthen the global search ability particles. The experimental results show that the RAQPSO algorithm greatly save the task completion time, and keep a good computing nodes load balancing.