一种粒子群结合遗传算法(PSO-GA)的云计算任务调度方法

Task Scheduling in Cloud Computing Based on PSO-GA Algorithm

针对云计算服务集群任务调度和负载平衡的优化问题,提出一种粒子群结合遗传算法(PSO-CA)的云计算任务调度方法。PSO-GA算法在遗传算法的基础上对种群进行分隔,用粒子群算法来构造变异算子,避免了变异算子的随机性和盲目性,很好地保持种群的多样性,克服了早熟现象。在Cloudsim平台进行模拟测试。实验结果表明,与同类算法相比,该调度方法能够缩短云计算下任务执行总时间,提高资源利用率。

Focusing on Cloud Computing services for task scheduling and load balancing cluster optimization problem, a PSO algorithm is proposed,combined with genetic algorithm （PSO-GA） in Cloud Computing task scheduling. Based on genetic algorithm separates populations, particle swarm algorithm is used to construct the mutation operator in PSO-GA algorithm. As a result, the mutation operator randomness and blindness are avoided, with population diversity maintained very well and premature phenomenon overcome. Extending the cloud computing emulator CloudSim platform to test the simulation, the re- sults show that PSO-GA can shorten the total cloud computin＇s task execution time and imr）rove resource utilization.