基于差分进化的BP神经网络在纱线质量预测中的应用  

Application of BP neural network based on differential evolution algorithm for the prediction of yarn quality

作　　者：曹成辉[1] 黄海涛[1] 王强[2] 

机构地区：[1]河南工程学院纺织工程系,河南郑州450007 [2]方圆标志认证集团山东有限公司,山东济南250000

出　　处：《河南工程学院学报（自然科学版）》2012年第3期1-5,共5页Journal of Henan University of Engineering：Natural Science Edition

摘　　要：为了提高BP神经网络在纱线质量预测上的精度,采用差分进化算法训练BP神经网络,利用差分进化算法的全局寻优能力优化BP神经网络的权值和阈值,建立了反映纱线质量的重要指标——单纱强度和条干CV%的神经网络预测模型.对真实数据的测试表明该算法效果良好,提高了BP神经网络算法的预测精度,能够为企业的纱线质量预测提供有效支持.In order to improve the prediction accuracy of the BP neural network in yam quality, the Differential Evolution （DE） algorithm was adopted to train the BP neural network, using the ability of global optimization of DE algorithm to optimize the weight val- ues and threshold values of BP neural network, and the prediction models were established for important indexes of yam quality： the single yam strength and the unevenness value CV. The experimental data showed that the algorithm proposed in this paper enhanced the prediction accuracy of the BP neural network algorithm, and could provide effective support for enterprises in the prediction of yam quality.

关 键 词：差分进化算法 BP神经网络 纱线质量 预测 

分 类 号：TS101.9[轻工技术与工程—纺织工程]