摘要
研究了运用模糊神经网络进行机械产品性能评价的原理和方法。着重说明了采用梯形公式的四元组形式表达网络中的隶属函数 ,对评价指标模糊化进行处理 ,以及采用改进的反向传播算法训练网络 ,得到合适的网络权值等几个方面。
It studies the evaluation for the performance of mechanical products based on fuzzy neural network, describes the fuzzy manipulation of evaluation targets with the membership function of trapezium formulae. This method uses the improved back propagation algorithm to get network coefficients. The application shows the process in performance evaluation of the hydraulic drifter.
