摘要
由于电子产品的设计周期越来越短,而对其降低成本和提高质量的要求却越来越高,电子线路的优化设计日益受到人们的关注。本文提出了一种实用的电子线路参数优化算法—均匀设计多项式拟合遗传优化算法。其主要思想是采用拟合的多元多项式代替电路仿真以指导遗传算法进行搜索,以便减少电路的仿真次数。并在数据采集过程中运用均匀设计方法,不仅能减少数据采样的数目,而且能使初始群体包含丰富的信息,更利于其收敛。电路优化实例表明,本算法对设计目标复杂以及参数较多电路的优化是有效的。
As it is required that electronic products are designed with shorter design period, lower cost and higher quality, circuit optimization has been paid increasing attention. In this paper, a practical approach to circuit parameter optimizationUniform Design Fitting Genetic Algorithm is proposed. The main idea is to replace circuit simulation with fitting of multivariate polynomial to guide the searching process of Genetic Algorithm as to reduce the number of times of circuit simulation. During datum sampling, Uniform Design is utilized, which can not only reduce the number of sampling, but also make initial population contain comprehensive information. This will be of benefit to the convergence of algorithm. Two circuit optimization experiments demonstrate that this approach is useful for those circuits with complicated objects having many design parameters.
出处
《电路与系统学报》
CSCD
2002年第3期115-118,共4页
Journal of Circuits and Systems
