食品科学 ›› 2008, Vol. 29 ›› Issue (7): 318-321.

• 分析检测 • 上一篇    下一篇

遗传算法与退火神经网络用于八种水溶性维生素液相色谱分离条件优化

陈昌云,李小华,马美华, 邵阳, 俞榕   

  1. 南京晓庄学院化学系; 江苏经贸职业技术学院工程技术系; 南京晓庄学院化学系 江苏南京210017; 江苏南京210007; 江苏南京210017
  • 出版日期:2008-07-15 发布日期:2011-07-28

Application of Genetic Algorithm and Annealing Neural Network to Optimize Separation Conditions for Eight Kinds of Water-soluble Vitamins by HPLC

 CHEN  Chang-Yun, LI  Xiao-Hua, MA  Mei-Hua, SHAO  Yang, YU  Rong   

  1. 1.Department of Chemistry, Nanjing Xiaozhuang University, Nanjing 210017, China; 2.Department of Engineering, Jiangsu Institue of Economic and Trade Technology, Nanjing 210007, China
  • Online:2008-07-15 Published:2011-07-28

摘要: 使用均匀试验设计法以甲醇在线性梯度展开时的初始浓度和线性梯度的斜率为优化参数,对八种水溶性VC、VB1、VB2、VB6、叶酸、VB12、烟酸和烟酰胺混合体系进行液相色谱分离条件优化。基于遗传算法和退火神经网络方法建立了有效的分离条件预测模型。对模型所预测的最佳分离条件进行试验,分离结果满意。将遗传算法和退火神经网络有机结合可有效地用于液相色谱分离条件优化。

关键词: 遗传算法, 退火神经网络, 水溶性维生素, 梯度分离条件优化

Abstract: Using the uniform test designs method, with the initial concentration of CH3 OH in the linear gradient and the slope of linear gradient as optimization parameters, the separation conditions for eight kinds of water-soluble vitamins were optimized, which are VC, VB1, VB2, VB6, folic acid, VB12, nicotinic acid and nicotinamide. Based on genetic algorithm and annealing artificial neural network, the predication model of efficient separation was established. Verification experiments were carried out with optimized separation conditions predicted by the model and the results are satisfying. Combination of the genetic algorithm and annealing neural network is available for the optimization of separation conditions of HPLC.

Key words: genetic algorithm, annealing neural network, water-soluble vitamin, optimization of gradient separation condition