【摘要】 利用遗传算法对神经网络的结构参数及初始权重进行了优化.得到优化的遗传神经网络GA-BP-ANN.将此网络应用于解析流动注射—光度分析法测得的邻、间、对硝基苯酚的重叠光谱,建立了不经分离同时测定邻、间、对硝基苯酚的新化学计量学方法.更多还原

【Abstract】 Flow injection analysis(FIA) is provided with high accuracy and high reproducibilily,Artificial Nural Network(ANN) that has selfadaptive and capability to solve problems of nonlinearity is a good mathematical method of multivariable calibration and resolution.Genetic Algorithm(GA) is a probability-searched global optimization method.In the search of GA application,two works have been done,first the global optimization capability of GA was used to optimize the structure parameters and connection weights of the artificial network respectively,and the optimized GA-BP-ANN has been applied to resolve the overlapped spectrum of the isomeoic compounds of nitrophenols which come from FIA.A new Chemometric method has been developed for simultaneous determination of onitrophenol,metanitrophenol and pnitrohenol without separation.更多还原