Abstract: The classification model of BP neural networks is put forward, which using electronic nose to acquire odor datum from four kinds of Chinese Liquors, aims at the research of Chinese liquors identification electronic nose. But the BP algorithm of neural network commonly used has several disadvantages, such as the slow convergence speed, the optimization procedure getting easily stacked into the minimal value locally and network parameter must be decided by experiment and experience. This paper designs a recognition classifier of Genetic Algorithm-Wavelet Neural Network（GA-WNN）, which has global optimization capability of GA, non-linear approximation ability of wavelet and self-learning characteristic of neural network. The simulation results prove that it can improve the recognition accuracy and convergence rate, and the GA-WNN algorithm can be used in Chinese liquors identification electronic nose.

Key words: Chinese liquors identification, electronic nose, wavelet neural networks, genetic algorithm

摘要： 为研究不同品质白酒快速识别的电子鼻技术，利用自制的电子鼻采集四种白酒样品的气味数据，建立了BP神经网络分类模型。针对BP算法普遍存在的收敛速度慢、易陷入局部极小且网络参数需要人工设定的缺陷，提出一种将遗传算法的自适应全局优化搜索能力、小波分析的非线性逼近能力和BP算法自学习能力结合在一起的遗传小波神经网络白酒识别模型。仿真结果表明，与BP神经网络和小波神经网络相比，GA-WNN分类模型的收敛速度和分类准确率都得到了较大提高，可应用于白酒识别电子鼻。

关键词: 白酒识别, 电子鼻, 小波神经网络, 遗传算法