基于用户满意度的任务调度策略

Task Scheduling Policy Based on User Satisfaction

为了满足云环境中用户任务调度的不同需求,提出一种改进粒子群算法的任务调度策略。将用户对时间和费用的期望值作为动态适应度函数的加权值,同时在粒子群算法中引入遗传算法的交叉和变异操作,不仅避免了算法陷入局部最优还保持解的多样性,最终求出满足用户需求的任务调度。仿真实验结果表明,该策略能够减低任务的完成时间和执行费用,提高云计算服务质量,具有良好的实用性。

In order to meet different needs of the user task scheduling in cloud environment,an improved particle swarm optimization algorithm of task scheduling strategy is proposed.Making the user expectations of the time and expense of value as a dynamic adaptive weighted function value,at the same time,it employs crossover and mutation operation of genetic algorithm,not only avoiding the algorithm into a local optimum,but also maintaining the diversity of solutions,and ultimately obtaining task scheduling with user needing.Simulation experimental results show that this algorithm can reduce the task execution time and cost,improve the quality of cloud computing services and possesses excellent practicability.