基于时间负载均衡蚁群算法的云任务调度优化

针对云渲染系统中渲染节点与任务不匹配调度而带来的时间负载不均衡和耗时长的问题,提出一种基于时间负载均衡的任务调度方式来优化系统耗时的策略.该算法采用Min-min与Max-min相结合的思想,建立时间负载均衡模型进行前期迭代,将迭代结果作为蚁群算法的初始序列,并按照适应度规则计算出相应的初始信息素,同时通过单一变量法确定合理的参数,蚁群算法采用已有的初始资源和参数值进行后期迭代,根据标准量度自定义函数进行高效寻优,进而求得最终的任务调度序列.仿真结果表明,本策略既具有较高的搜索效率和较强的全局寻优能力,又能有效降低任务完成时间,且在时间负载均衡和寻优速度方面均显著优于蚁群算法和蚁群退火算法.

Task Scheduling and Optimization of Cloud Computing Based on Ant Colony Optimization of Time Load Balance

For cloud rendering systems,a task scheduling method of optimizing the system time-consuming strategy based on time load balancing is proposed to solve the problems of unbalanced time load and long time consuming caused by mismatching scheduling of rendering nodes and task.The idea of Min-min and Max-min algorithm is used to establish the time-load equilibrium model for the pre-iterative process.The pre-iterative result is used as the initial sequence of the ant colony optimization(ACO),and the corresponding initial pheromone is calculated according to the fitness rule.Simultaneously,the ACO uses the initial resources and the parameter values obtained by a single variable method to perform later iteration,and applies it to the standard measure customized function to seek the optimal sequence efficiently,then obtains the final task scheduling sequence.The simulation results show that the strategy not only has high search efficiency and strong global optimization ability,but also can effectively reduce the task completion time.Simultaneously,it is superior to ant colony optimization and ant colony annealing algorithm in terms of time load balancing and optimization speed.