基于锯齿型微粒群方法的放大器参数优化

The Designing of Optimal Amplifier Based on Particle Swarm Optimization

针对放大器电路中多种参数的同时优化问题,对微粒群算法进行改进,提出一种锯齿型动态群体数量的微粒群方法,并将其用于放大器电路参数的优化设计.改进的锯齿型微粒群方法中,群体数量在每个固定的进化阶段线性减少,当群体数量减少到给定最小值时,利用交叉操作产生新个体对群体进行补充.此方法不但减少算法的运算量,而且减少因随机操作产生新个体导致平均适应度振荡的弱点.为验证方法的有效性,以放大器参数优化问题为研究对象,在高频条件下建立非线性电路及非线性器件的数学等效模型,利用不同的微粒群方法实现模拟电路的自动设计,对比实验验证方法的有效性.

In order to design the optimal parameters of amplifier circuit simultaneously,a new saw-toothed particle swarm optimization（PSO） is proposed based on the improvement of conventional PSO.The new method is used to realize global optimization for parameters of amplifier.In the given method,the number of swarm decreased linearly on each period of evolution,crossover operator is used to generate new individual for current swarm when the number is arrived the setting minimal value.The computation cost of conventional PSO is decreased and the surging of average fitness for PSO with random generating new individual is overcome.In order to indicate the effectiveness of the given method,The optimization of parameters for amplifier is studied in the simulation experiment,the equivalent mathematical models of nonlinear circuit and components are built on condition that the frequency is high,different PSOs are used to optimize the parameters of analogy circuit automatically.The effectiveness of the method is validated with comparing simulations.