基于AHP-BP神经网络的航空材料供应商选择

Selection of aeronautical material suppliers based on AHP-BP neural networks

根据适航规章要求,提出了航空材料供应商选择的特色指标,构建了航空材料供应商选择的AHP-BP神经网络模型,通过AHP确定各指标权重,再结合BP神经网络,从训练数据中提取隐含的知识和规律,能够方便地用于新供应商的选择。该模型求解算法为动量梯度下降反向传播算法,具有良好的可扩展性,从而增加了评价的动态性。算例验证结果表明,将AHP-BP神经网络模型用于航空材料供应商选择具有较强的实用性。

According to the requirements of airworthiness regulations, a special index system for the selection of aeronautical material suppliers is proposed with the establishment of AHP-BP neural network model. The weight of each index is determined through AHP, and then combined with BP neural networks, the implicit knowledge and discipline from the training data is extracted, so that it can be used for new suppliers conveniently. The momentum gradient descent back-propagation algorithm is used in the model with well extensibility and increasing dynamic nature of the evaluation. Verification shows that AHP-BP neural network model is with strong practicability for the selection of aeronautical material suppliers.