A Circuit Generating Mechanism with Evolutionary Programming for Improving the Diversity of Circuit Topology in Population-Based Analog Circuit Design
A Circuit Generating Mechanism with Evolutionary Programming for Improving the Diversity of Circuit Topology in Population-Based Analog Circuit Design
复制标题
DOI：
10.1007/978-3-642-38703-6_63
复制
发表时间：
2013-06
期刊：
影响因子：
--
通讯作者：
Mei Xue;Jingsong He
中科院分区：
文献类型：
--
作者：
Mei Xue;Jingsong He
提出了一种基于电路网表中连接点引导的模拟电路生成机制。利用该机制，初始电路拓扑可以是一个随机网表，进化操作可以直接在连接点上执行。此外，图论的知识被引入到评估电路结构的多样性程度。实验结果表明，该机制有利于提高种群拓扑结构的多样性。在没有鲁棒性进化机制的情况下，种群拓扑结构的多样性可以提高种群的容错性。
This paper presents an analog circuit generating mechanism based on connecting point guidance existing in circuit netlist. With the proposed mechanism, the initial circuit topology can be a random netlist, and the evolutionary operation can be executed directly on connecting point. Also, the knowledge of graph theory is introduced for evaluating the degree of diversity of circuit structures. Experimental results show that the proposed mechanism is beneficial to improve the diversity of topology in population. In the case of no robustness evolution mechanism, the diversity of topology in population can improve the fault tolerance of population.