基于遗传算法与共轭梯度法结合的四维同化方法  

A Four-Dimension Assimilation Scheme Based on Genetic Algorithm and Conjugate Gradient Method

作　　者：白晨[1] 吴诚鸥[1] 吴令云[2] 

机构地区：[1]南京信息工程大学数学系,江苏南京210044 [2]南京审计学院信息科学学院,江苏南京210029

出　　处：《南京气象学院学报》2006年第6期850-854,共5页Journal of Nanjing Institute of Meteorology

基　　金：全国统计科研计划项目(LX2005-Y18)

摘　　要：对气象资料的四维同化,采用共轭梯度算法和遗传算法,可以使两种方法相互补充,实验证明两者的结合能够增强搜索能力和收敛速度。在编写此程序时,采用标准化编程方法,使程序能够接近全自动化,从而缓解了编程的繁重劳动,减少程序中出现的错误。In four-dimension assimilation, combining conjugate gradient metnod and genetic algorithms （DA） could remedy the shortcoming of each other. It is proved in the experiments of this papar that the scheme not only could strengthen searching ability but also could accelerate the convergent speed of optimization. We adopt standard program in programming, therefore the assimilation program is almost automatic, so we could be eased from the hard programming work and avoid most of the errors.

关 键 词：遗传算法 共轭梯度 四维同化 

分 类 号：O231[理学—运筹学与控制论]