基于改进粒子群算法的云计算调度策略

Cloud Computing Scheduling Strategy Based on Improved Particle Swarm Optimization Algorithm

云计算调度策略是一种将海量计算任务分配到各个计算资源上的模型。粒子群优化算法作为一种随机全局搜索算法在云计算中广泛应用,然而其仍存在参数依赖人为设定、前期全局搜索能力不足、后期收敛速度缓慢等问题。针对上述问题,提出了基于模拟退火策略与自适应权重策略的优化方法,旨在提高粒子群调度算法的自我优化能力。通过引入上述策略,对粒子群算法中的学习因子、速度增长等参数进行实时控制,完成算法迭代优化。实验表明,在CloudSim平台中该优化策略增强粒子群中每个粒子的学习能力,与优化前相比执行速度更快,精确度更高。

Cloud computing scheduling strategy is a model that allocates massive computing tasks to various computing resources. As a random global search algorithm,particle swarm optimization algorithm is widely used in cloud computing.However,there are some problems,such as parameter dependence on artificial setting,insufficient global search ability in the early stage,slow convergence speed in the later stage and so on. To solve the above problems,an optimization method based on simulated annealing strategy and adaptive weight strategy is proposed to improve the self-optimization ability of particle swarm optimization algorithm. By introducing the above strategy,the learning factors,speed growth and other parameters in particle swarm optimization algorithm are controlled in real time,so as to complete the iterative optimization of the algorithm. Experiments show that in Cloud Sim platform,the learning ability of each particle in particle swarm optimization is enhanced by the optimization strategy,with faster execution speed and higher accuracy than before optimization.