摘要
针对传统的网格任务调度算法存在的缺陷,提出了用分层遗传算法来实现对网格任务调度策略的优化.在构造分层遗传算法时引入了SGA,AGA和CHC算法.SGA采用基本的遗传操作,保证了种群的多样性;AGA对交叉概率和变异概率的动态调整,保证了遗传算法的收敛性;CHC算法强调优良个体的保留,加快了遗传算法的收敛速度;分层遗传算法在吸收了这3种算法优点的基础上进行优化.实验结果表明,分层遗传算法在结果精度和收敛速度上都较其他算法有较大程度的提高.
A hierarchical genetic algorithm is proposed to optimize the grid task scheduling strategy. Standard genetic algorithm (SGA), adaptive genetic algorithm (AGA) and CHC algorithm are brought into the hierarchical genetic algorithm. Simple genetic operators are used in SGA to improve the population’s diversity; dynamic cross rate and mutation rate are used in AGA to ensure convergence, and excellent members are reserved in CHC algorithm to increase the speed. The hierarchical genetic algorithm adopts the benefits of the three kinds of algorithms and make some ameliorations to enhance the efficiency. The experiment has proved that quality and efficiency are improved markedly in comparison with other algorithms.
出处
《计算机研究与发展》
EI
CSCD
北大核心
2008年第z1期35-39,共5页
Journal of Computer Research and Development
基金
国家自然科学基金项目(90612016)
网络计算的作业调度方法研究基金项目(60473095)
关键词
分层遗传算法
网格任务调度
自适应遗传算法
CHC算法
基本遗传算法
hierarchical genetic algorithm
grid task scheduling
adaptive genetic algorithm
CHC algorithm, standard genetic algorithm
