同其他模拟技术相比，在精度相同的情况下，神经网络响应面法可以大大减少模拟时间。
Compared with other simulation technique, neural network response surface method can greatly reduce the simulation time at the same precision level.
首先通过一系列样本点有限元试验结果建立目标函数和设计变量的神经网络响应面模型。
A response surface using artificial neural network is formed on the ground of initial finite element experiments.
经过神经网络响应面法的检验，证明TAA方法是一种实用而又有效的风险概率计算方法。
Verified by network response surface, TAA was improved to be an applied and valid method of calculating the risk occurrence probability.
分析了通过神经网络来实现响应面模型构建的方法。
The construction of response surface models by the artificial neural network is analyzed.
最后通过GA - BP神经网络与拉丁超立方抽样法相结合构建了可控拉深筋主要影响因子h1和H2与极限拉深深度之间的响应面。
Eventually, the response surfaces composed of the CD main influence factor H1, H2 and limit drawing depth are established by the combination of GA-BP neural network and Latin Hypercube.
通过线性回归和相似因子法比较人工神经网络和基于二元二项式的响应面法的预测能力。
The regression equations and similar factors were used to compare the predicability of the ANN and response surface method based on the binomial equations.
通过线性回归和相似因子法比较人工神经网络和基于二元二项式的响应面法的预测能力。
The regression equations and similar factors were used to compare the predicability of the ANN and response surface method based on the binomial equations.
应用推荐