实验基于资源预先分类的云计算任务调度算法

Task Scheduling Algorithm Based on Improved Particle Swarm Optimization Algorithm in Cloud Computing Environment

研究云计算中如何合理分配计算资源及有效调度任务运行的问题,针对云计算中现有的任务调度算法只追求任务最短完成时间,而没有从云计算中资源负载均衡的角度考虑,容易导致云计算中一些资源负载过大而另一些资源闲置的现象,为改善最短完成时间而不能很好兼顾负载均衡的问题,提出了一种利用资源预先分类的具有双适应度的粒子群优化调度算法,将云计算资源属性考虑在内,作为另一适应度函数对任务进行调度。通过改进算法调度产生的结果不仅能使任务完成所需时间较短,而且系统资源利用率较高,兼顾了执行时间最小和负载均衡。仿真结果表明,在相同的条件下,改进算法优于传统的粒子群优化算法,为云计算有效地优化调度提供了依据。

Research the problem of that rational allocation of computing resources and effective scheduling of tasks in the cloud. Existing task scheduling algorithms for cloud computing do not take into account the load balan- cing problem for the pursuit of the shortest completion time. To solve this problem, a double - fitness particle swarm optimization（DFPSO） based on resource pre -classification was proposed in this paper. In the new algorithm, the properties of the cloud computing resources were considered, which is regarded as another fitness function of task scheduling. The results generated by this algorithm not only make the task completion time shorter, but also have a higher utilization of system resources, which takes into account the minimum execution time and load balancing. The simulation shows that DFPSO is an efficient task scheduling algorithm in the cloud computing by contrast with the conventional particle swarm optimization（PSO）.