人工神经网络结合遗传算法在焙炒大米过程中的应用

Application of Genetic Algorithm Associated with Artificial Neural Network Techniques for Optimization of Rice Roasting Process

焙炒是一种全新的使原料淀粉糊化的方法.它用热风替代水蒸气,在高温,短时间的条件下处理原料米,具有无废水污染,容易保存等优点,为食用酒酿制过程中的一种新型技术.通过对焙炒大米的3个指标:糊化率、脂肪含量和氨基氮进行测定,利用人工神经网络技术ANN对上述性能指标和操作参数的数据进行训练学习,得到可以描述焙炒过程操作条件和性能指标之间关系的模型.在所得模型的基础上,利用遗传算法GA对大米的焙炒条件实施优化,对未参与ANN建模的数据进行评价和比较,结果发现,结合使用ANN和GA,能够比较准确地预测对应于期望指标的操作条件,预测结果与实验数据吻合.

Rice roasting process is a novel technique for raw starch gelatinization. It uses heated air to replace water steam for processing rice under the conditions of high temperature and short time, and is a new technique for making rice wine with the characteristics of easy storage and zero water pollution. This study measured three major process performance index, starch α-ratio, total fat and amino nitrogen contents of the roasted rice under various operating conditions, then the relationship between the performance index and the corresponding operation variables were modeled by artificial neural network (ANN) by learning and training the corresponding data sets. Based on the models obtained, genetic algorithm was used to optimize the rice roasting process, which is to search the optimal operation variables corresponding to the expected performance index. The results showed that the combination of artificial neural network and genetic algorithms could accurately estimate the operating variables for the expected performance index, by evaluating several data sets unused for the ANN models' learning and training.