Abstract:

Although regression analysis can predict some drape indicators, they have problems such as low prediction accuracy and the inability to calculate some indicators. To overcome these issues, this study proposes a new method using genetic algorithm to optimize BP neural network (GA-BP neural network) to improve the prediction accuracy of real fabric drape. In this study, we designed a GA-BP neural network model, selected 100 pure cotton woven fabric samples from the fabric database, including 80 training samples, 10 test samples, and 10 validation samples, used the genetic algorithm to optimize the parameters of the neural network, and used correlation analysis to optimize sample input parameters to improve the prediction performance of the model. The results of the drape coefficient prediction for the 10 test samples show that compared with the traditional BP neural network, the average absolute percentage error of the BP neural network optimized by the genetic algorithm decreased from 12.74% to 7.03%. Furthermore, we used an empirical equation to identify error cycles and concluded that the optimal number of hidden layer nodes is 9. This study indicates that the GA-BP neural network can effectively improve the accuracy of fabric drape prediction and has important application value for the virtualization of fabric drape performance.

Key words: drape coefficient, fabric database, neural network, genetic algorithm