计算机工程与应用 ›› 2017, Vol. 53 ›› Issue (20): 148-153.DOI: 10.3778/j.issn.1002-8331.1703-0486

• 模式识别与人工智能 • 上一篇    下一篇

基于结构矩阵的电路网表编码方法的研究

孙照兴,何劲松   

  1. 中国科学技术大学 信息科学技术学院,合肥 230027
  • 出版日期:2017-10-15 发布日期:2017-10-31

Research on method of circuit netlist coding based on structural matrix

SUN Zhaoxing, HE Jinsong   

  1. School of Information Science and Technology, University of Science and Technology of China, Hefei 230027, China
  • Online:2017-10-15 Published:2017-10-31

摘要: 在模拟电路演化领域,电路知识表示是首要解决的问题。网表编码操作简单,对于拓扑结构没有限制,但是网表编码会在种群初始化和遗传操作过程中产生大量的非法电路个体。为解决这一问题,提出结构矩阵,并总结出合法电路结构矩阵所具有的性质,以结构矩阵为规范设计出合适的种群初始化步骤和能够用于网表编码的结构交叉算子,通过演化来验证效果,实验结果显示该方法能够较好地解决网表编码所存在的问题。

关键词: 模拟电路演化, 结构矩阵, 交叉算子, 网表编码

Abstract: Circuit knowledge representation is the primary problem in the analog circuit evolution. The netlist coding is brief and has no limit on the topology, however, it generates a large number of illegal circuits during population initialization and genetic operations. This paper presents a structural matrix approach to solve the problem. The approach first summarizes the nature of the legal structure matrix, then uses the properties design population initialization steps and appropriate crossover operator. Proof of concept is given by two sets of evolutionary experiments. The results show that the proposed method can solve the problem of netlist coding.

Key words: analog circuit evolution, structure matrix, crossover operator, netlist coding