不确定执行时间的云计算资源调度

Resource Scheduling with Uncertain Execution Time in Cloud Computing

针对执行时间不确定情况下的云计算资源调度问题,基于模糊规划理论建立了时间-成本约束条件下的模糊云资源调度模型,使用三角模糊数表示不确定的任务执行时间,以最小化评价函数的平均值和不确定度作为调度目标。提出一种改进的混沌蚁群算法对模型进行求解,算法引入精英策略优化了信息素的更新,采用折叠次数无穷大的混沌映射进行混沌搜索,并设计了自适应混沌扰动机制以增强算法的全局搜索能力。在Cloudsim平台上用仿真数值实例对模型和算法进行验证,证明了模型的可靠性,实验结果表明改进算法在收敛速度、求解能力和负载均衡上均有较好的性能。

For the problem of cloud computing resource scheduling,based on the fuzzy programming theory,a fuzzy cloud resource scheduling model under time-cost constraint was set up,the uncertain execution time of tasks is represented by the triangular fuzzy number,and the target is to minimize the average value and standard deviation of the evaluation function.An improved chaotic ant colony algorithm was proposed to solve the model,the elitist strategy is introduced to optimize the pheromone updating,a chaotic mapping with infinite folding times is used for chaotic search,and the adaptive chaotic disturbance mechanism is designed to enhance the global searching ability.The model and algorithm were tested on the Cloudsim platform,the reliability of the model was proved,and the experimental results showed that the proposed algorithm had better performance in convergence speed,solution ability and load balance.