一种基于改进遗传算法的网格调度算法  

A Grid Scheduling Algorithm Based on Improved Genetic Algorithm

作　　者：杨新年[1] 刘锋[1] 田铁刚[1] 汤泰青[1] 李晓艳[1] YANG Xinnian;LIU Feng;TIAN Tiegang;TANG Taiqing;LI Xiaoyan(Heilongjiang University of Technology,Jixi 158100)

机构地区：[1]黑龙江工业学院,鸡西158100

出　　处：《计算机与数字工程》2018年第9期1786-1789,1795,共5页Computer & Digital Engineering

基　　金：2016年黑龙江省自然科学基金(面上项目)"网格计算中任务调度算法与资源发现模型的研究"(编号:F2016038)资助

摘　　要：为提高网格调度的高效性,针对在种群规模大、迭代数多时的情况下,传统遗传算法性能较低的问题,提出一种改进的遗传算法。该算法通过结合Min-min算法来生成初始种群,利用双适应度函数的方式来改进选择、交叉、变异的过程,同时改进收敛判断函数,以使进化过程中既保证种群中的高质量个体有较高的比例同时保证种群的多样性,最终算法能够尽快收敛同时又可得到最优解。实验结果表明该算法的性能在规模较大的情况下比传统算法要优秀。In view of the population size is larger,number of iterations is large,the traditional genetic algorithm performance is lower,in order to improve the efficiency of grid scheduling,an improved genetic algorithm is put forward. The algorithm generates the initial population by combining the Min-min algorithm,uses dual fitness function to improve the process of evolution,improvs convergence judging function at the same time to make the algorithm can not only ensure the diversity of population,and the convergence as soon as possible. The experimental results show that the performance of the algorithm in the case of a larger more excellent than the traditional algorithm.

关 键 词：网格计算 遗传算法 网格调度 收敛性 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]