一种以模糊数学为基础的布局方法

A NEW PLACEMENT ALGORITHM BASED ON FUZZY SET THEORY

本文从布局问题相关联的二次分配问题的连接矩阵出发,应用模糊集合论推出了一个具有自反性、对称性和最大积传递性的模糊相似关系.根据这个模糊相似关系和经过修改的模糊c-划分聚类分析方法,本文提出了一种新的自动布局算法.实验结果表明,这个算法优于许多常用的布局方法,而且这个算法具有快速、灵活的特点,所以适用于VLSI设计.

The connection matrix of the associated quadratic problem derived from a placement problem is modified using fuzzy set theory. The modified connection matrix, called fuzzy similarity relation, is self-reflective, symmetric and max-product transitive. A new placement algorithm based on the fuzzy similarity relation and a modified fuzzy c-means clustering method are presented. Experimental results show that the performance of the method is superior Co a number of existing placement algorithms.