一种基于物理模型与遗传算法的平面螺旋电感的优化技术

针对 CMOS RF平面螺旋电感的物理模型 ,提出了一种使用遗传算法优化电感设计参数的优化技术 .应用这种技术 ,人们无需再通过等 Q值线来求取 Q值最佳的电感设计方案 ,因为等 Q值线的计算代价很高 ,而且一旦工艺条件改变 ,所有的等 Q值线都要重新计算 ,效率很低 .而这里提出的优化技术在计算时并没有将所有的电感版图设计方案都计算一遍 ,因而计算效率大大提高 .计算结果与相应设计方案的测量参数比较 ,只有约 5 %的计算误差 .计算结果与相同条件下的等 Q值线的最佳设计方案比较 ,这种优化技术总能得出与之基本一致的最佳方案 .事实上 ,由于这种优化技术

An Optimization Technique for Planar Spiral Inductor Based on the Inductor's Physical Model and Genetic Algorithm

An efficient optimization technique for CMOS RF planar spiral inductor is presented on the basis of the physical model of an inductor as well as the genetic algorithm.With it,it is unnecessary to design an optimal high Q inductor by means of the contour plots of Q ,which is high cost and time consuming as we all know.Via the new technique,the computation efficiency is improved greatly,because we haven't to compute each design scheme.The calculated result is in agreement with the measured one,with the error within 5%.By using the technique,the same optimal design can be obtained under the same condition as that obtained by using contour plots.As a matter of fact,because of the flexibility of our technique,the acturally best design can be only obtained by using this technique instead of by the contour plots of Q .