基于优化DPSO算法的云平台任务调度研究

Research on Task Scheduling Algorithm Under Cloud Platform Based on Optimized DPSO

为提升离散粒子群优化算法(discrete PSO,DPSO)的全局收敛性和收敛效率,提出一种基于适应值的分段自适应惯性权重.根据粒子在空间搜索过程中适应度值的大小,将粒子的搜索性能分为4个状态区,粒子处于不同的状态区,拥有不同的惯性权重值.当粒子当前的适应值接近粒子群中最优粒子的适应值时,应赋予粒子较小的惯性权重值,反之,应赋予粒子较大的惯性权重值.通过动态调整粒子所处各个阶段的搜索状态,来加速粒子向全局最优解收敛.提升DPSO算法的全局搜索性能,并将优化的DPSO算法应用于云平台的任务调度.仿真实验表明,优化后的DPSO算法具有高效的全局搜索性能,能快速地为云平台提供最佳任务调度策略.

In order to improve the global convergence and convergence efficiency of the DPSO algorithm,a piecewise adaptive inertia weight was proposed based on the adaptive value in this paper.According to the size of the particle in the space search process,the search performance of the particle was divided into four state regions and in different state area the particles had different inertia weight value.When the current adaptive value of the particle was close to the adaptive value of the best particle in the particle swarm,the particle would be given a smaller inertia weight value.Otherwise,the bigger inertia weight value would be given to the particle.By dynamically adjusting the searching state of particles at all stages,the particle could be accelerated to converge to the global optimal solution.We improved the global search performance of DPSO algorithm and applied the optimized DPSO algorithm to task scheduling under cloud platform.Simulation results showed that the optimized DPSO algorithm had high efficiency in global search and provided optimal scheduling strategy for cloud platform quickly.