基于MGA-PSO的云计算多目标任务调度

MULTI-OBJECTIVE TASK SCHEDULING OF CLOUD COMPUTING BASED ON MGA-PSO

为了提高云计算任务调度的效率,将微生物遗传算法(MGA)和改进的粒子群算法(PSO)融合成MGA-PSO算法用于云计算任务调度。综合任务完工时间、任务执行成本及虚拟机负载均衡三个目标构造适应度函数,以此寻找任务调度的最优解;对粒子群算法进行改进,使用动态惯性权重策略以提高算法的自适应搜索能力;在任务调度前期使用MGA算法缩小求解空间,在任务调度后期使用改进的PSO快速收敛到最优解。仿真实验表明:与其他三种算法相比,该算法有较快的收敛速度和较强的寻优能力;在云计算任务调度中,不仅能减少任务完工时间和执行成本,还能优化虚拟机的负载。

In order to improve the scheduling efficiency of cloud computing tasks,microbial genetic algorithm(MGA)and the improved particle swarm optimization(PSO)are merged into MGA-PSO for cloud computing task scheduling.The three objectives of the task completion time,task execution cost and virtual machine load balancing construct the fitness function to find the optimal solution of the task scheduling.The PSO was improved,and the dynamic inertia weight strategy was used to improve the adaptive search ability of the algorithm.In the early stage of task scheduling,the MGA was adopted to reduce the solution space,and the improved PSO was used to rapidly converge to the optimal solution in the later stage of task scheduling.The simulation experiments show that compared with the other three algorithms,this algorithm has faster convergence speed and stronger optimization ability.In cloud computing task scheduling,it can not only reduce task completion time and execution cost,but also optimize the load in the virtual machine.