关键生产参数挤出式乳胶丝质量建模与应用

Quality modeling and application of extruded latex thread based on key parameters

为满足多工序挤出式乳胶丝高效稳定生产需要,基于挤出式乳胶丝关键生产参数,采用多重多元回归、广义回归神经网络(generalized regression neural network,GRNN)与量子遗传算法(quantum genetic algorithm,QGA)分别建立多重多元回归模型、GRNN模型与QGA-GRNN模型,并以均方误差(mean squared error,MSE)作为评价指标在应用中比较3个模型预测效果。应用结果表明:QGA-GRNN模型在应用中能取得更小的MSE,预测能力优于其他两个模型,适用于挤出式乳胶丝产品质量预测。

In order to meet the need of the efficient and stable production of extruded latex threadof multiple procedures, multivariate regression model, GRNN model and QGA-GRNN model,based on the key parameters, were established by using multivariate regression, generalizedregression neural network (GRNN) and quantum genetic algorithm (QGA). The mean square error(MSE) is applied as the evaluation index to compare 3 models. The application result indicatesthat QGA-GRNN model obtains smaller MSE, and its forecasting capacity is superior to the other two models. It is therefore appropriate for extruded Latex latex thread product quality prediction.