基于模糊逻辑的器件与电路建模技术

器件与电路的模型是进行CAD的基础,建模在数学上实质是函数逼近问题。模糊逻辑具有多层前向神经网络类似的万能函数逼近特性,本文介绍了利用这一技术进行器件与电路建模的一些尝试。由于模糊逻辑能自然地表达与问题有关的知识与经验,因此有可能取得较神经网络更好的结果。文章介绍了用模糊逻辑建模的过程,提出了用数论网格的样本点生成方法,最后用实例说明了用于器件与电路建模的效果。

Circuits and Devices Modeling by Means of Fuzzy Logic

Circuit and device modeling play the fundamental role in the CAD of electronic circuits. Mathematically, modeling is essentially a problem of multi-dimensional function approximation. In this paper, a method for device and circuit modeling by means of the fuzzy logic is presented. Its basic principle is to make use of the universal function approximation capability of feed-forward neural networks. The modeling procedure is described. A new sampling scheme for training based on the number theoretic lattice is proposed. Illustrative examples are given to verify the effectiveness of the proposed method.