网格任务调度问题求解的萤火虫算法  

Glowworm swarm algorithm for solving grid task scheduling optimization problem

作　　者：张拓[1,2] 王建平[1] 

机构地区：[1]合肥工业大学电气与自动化工程学院,安徽合肥230009 [2]淮北职业技术学院建筑工程系,安徽淮北235000

出　　处：《重庆邮电大学学报（自然科学版）》2015年第5期654-659,共6页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：安徽省"十二五"科技攻关计划项目(11010402183)~~

摘　　要：为了获得更优的网格任务调度方案,针对网格环境特点以及标准萤火虫算法存在的不足,提出了一种基于改进萤火虫算法的网格任务调度优化模型。对网格任务调度优化问题进行了分析,建立了网格任务调度的数学模型,引入非均匀变异算子和自适应步长的搜索策略,加快算法的求解速度和精度,并将改进萤火虫算法用于网格任务调度问题求解,通过萤火虫之间的信息共享和交流找到网格任务调度最优方案,采用仿真对比实验对其有效性和优越性进行测试。结果表明,相对于其他网格任务调度优化算法,改进萤火虫算法可以快速、准确地找到网格任务的最优调度方案,提高了计算资源的利用率,保证了网格系统负载均衡,尤其对于大规模网格任务调度问题,具有更加明显的优势。In order to obtain better grid task scheduling scheme,according to the characteristics of grid environment and the shortcomings of traditional glowworm swarm optimization algorithm,this paper presents a grid task scheduling optimization model based on improved glowworm swarm algorithm. First of all,the optimization problem of grid task scheduling is analyzed and the mathematical model of grid task scheduling is established,and then aiming at the shortcomings of standard glowworm swarm algorithm,non uniform mutation operator and adaptive step size search strategies are introduced to solve fastening the speed and improve accuracy and the improved the firefly algorithm is used to solve the problem of grid task scheduling. The optimal scheduling scheme of grid task is found by firefly between the information sharing and exchang. Finally the simulation experiment is used to test the validity and superiority. The simulation results show that,compared with other optimization algorithms for grid task scheduling,the improving glowworm swarm optimization algorithm can quickly and accurately find the optimal scheduling of grid tasks,improving the utilization of computing resources,and ensuring the load average grid system,which has more obvious advantage especially for the large scale grid task scheduling problem.

关 键 词：网格任务 萤火虫算法 调度模型 自适应步长 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]