免疫遗传神经网络算法的纱线原料性能反演

Property Inversion of Yarn Raw Material for Immune Genetic Neural Network Algorithm

探讨基于免疫遗传神经网络算法的纱线原料性能参数反演。以BP神经网络建立纱线原料性能参数正演模型,同时使用免疫遗传算法优化该网络的权值和阈值,以提高网络的预测精度和速度,在此基础上,以纱线强力值为对象再次通过免疫遗传算法构建反演模型,并对反演参数进行求解。通过真实数据训练仿真,反演精度达到了95%,验证了该方法的可行性与有效性。认为:该方法可为纺织企业的配棉工作和工艺设计提供有效的理论指导。

The parameter inversion of yarn raw material properties based on immune genetic neural network algorithm was discussed.BP neural network was used to establish the parameter forward model of yarn raw material properties.Meanwhile,immune genetic algorithm was adopted to optimize weight and threshold value of the network to increase prediction accuracy and speed of the network.Based on this,yarn strength value was taken as object and then the inverse model was built with immune genetic algorithm again to solve inversion parameter.Through training and simulation of truthful data,the inversion accuracy was reached 95%.Feasibility and effectiveness of the method were verified.It is considered that the method can provide effective theoretical direction for textile enterprises on cotton assorting and processing design.