论文部分内容阅读
本文介绍了一种新的工艺研究途径──神经网络反向传播算法（BP算法），并对其网络学习收敛速度的加快进行了讨论，取得了较好的效果；利用改进的神经网络BP算法，我们通过建立IC表面钝化工艺的神经网络模型，对IC表面钝化工艺进行了计算机模拟，精确地预报了实验结果并得到了相关工艺条件与钝化介质膜特性的关系曲线．所编写的神经网络应用程序已用C语言在计算机上得到了实现．
In this paper, a new approach to technology research ─ ─ neural network backpropagation algorithm (BP algorithm) is introduced, and its convergence speed of network learning is discussed, and achieved good results. By using improved BP neural network algorithm , We set up a neural network model of the IC surface passivation process to simulate the IC surface passivation process and accurately predict the experimental results and get the curve of the relevant process conditions and passivation dielectric film characteristics. The prepared neural network application has been implemented on a computer using C language.