摘要
提出了一种复杂样品线性梯度分离条件的快速优化方法。先通过4次线性梯度的初始实验,然后利用Excel规划求解工具获得准确的溶质保留方程系数。在此基础上利用网格搜索完成线性梯度的参数优化。整个优化策略通过对中药金银花提取液中各组分的分离得以验证。同时比较了网格搜索、遗传算法、遗传算法与规划求解联用的3种寻优算法。结果表明:3种方法优化结果接近,但运算时间有所差异。
A novel method for the fast optimization of gradient condition for herbal medicine＇s chromatographic fingerprint is proposed. The retention equation＇s coefficients of components can be obtained using the Microsoft Excel Solver by running four linear gradient elutions and then the gradient parameters were optimized by grid search algorithm. The whole optimization strategy was validated by the separation of Chinese traditional medicine Flos Lonicerae extracts. Moreover, three optimization algorithms, such as grid search, genetic algorithm, and the combination of genetic algorithm and the Microsoft Excel Solver, were compared for solving the problem of parameters optimization. The results showed the results using the three different optimization method were very close but the calculation time was obviously different.
出处
《分析化学》
SCIE
EI
CAS
CSCD
北大核心
2007年第6期839-844,共6页
Chinese Journal of Analytical Chemistry
关键词
中药
指纹图谱
线性梯度
优化
Herbal medicine, chromatographic fingerprint, gradient elution, optimization
