摘要
采用计算模拟的方式研究了遗传算法和线性回归法反演气溶胶粒径分布的准确性。发现遗传算法能在一定程度上降低Mie矩阵向量间多重共线性给反演带来的误差,其模拟结果优于普通线性回归法。
The accuracy of retrieving the aerosol size-distribution by using the Genetic Algorithm and linear regression is studied through computer simulations. It is found that the error comes from the multi-collinearity among the vectors of Mie matrix could be reduced to some extent when using Genetic Algorithm, and the results of Genetic Algorithm are better than those of the linear regression.
出处
《光散射学报》
北大核心
2009年第4期332-334,共3页
The Journal of Light Scattering
关键词
粒径分布
遗传算法
线性回归
size-distribution
Genetic Algorithm
linear regression
