基于思维进化优化灰色神经网络在真空玻璃需求中的研究

Research on forecasting vacuum glazing demand based on grey neural network with MEC algorithm

由于真空玻璃需求受到多重因素的影响,传统的数据模型无法准确找寻订单变化规律,导致预测精度较低。为了提高预测精度,结合灰色神经网络模型,利用思维进化算法优化灰色神经网络,确定灰色神经网络的最优初始参数。分别利用灰色神经网络(GNN)模型和思维进化-灰色神经网络(MEC-GNN)模型,用训练好的网络预测某真空玻璃制造商订单,预测结果表明:改进的MEC-GNN模型明显地提高了预测结果的精度。

The forecast precision gets low, because the demand of vacuum glazing is affected by multiple factors and the traditional models can not accurately find the rule. In order to improve the prediction accuracy, the optimal initial parameters are determined by using the grey neural network model and MEC algorithm. Forecast the vacuum glazing manufacturer orders with network pre- trained by using the grey neural network（GNN）model and mind evolutionary computation- grey neural network（MEC-GNN）models respectively. The forecast results show that the MEC-GNN model can improve the prediction accuracy.