基于混合遗传算法的去耦电容网络设计

Design of decoupling capacitor network based on hybrid GA

结合基于目标阻抗的设计方法,将混合遗传算法应用于PDN(电源分配网络)的去耦电容网络设计,对所需去耦电容器的种类和数目进行优化计算,同时与标准遗传算法和传统设计方法进行对比,借助Cadence电源完整性仿真工具验证设计结果的有效性和合理性。避免了传统设计需要借助EDA工具进行反复修改和验证的繁琐过程,同时减少了设计冗余,节省了PCB布局布线空间,为高频电路的PDN设计提供参考。

Combined with the design method based on target impedance, Hybrid CA （Genetic Algorithms） was applied to the PDN （Power Distribution Net） decoupling capacitor network design to optimize the types and the required number of decoupling capacitors. The result was compared to that of the standard GA and the traditional design method, and then verified by the power integrity simulation tools of Cadence. It avoids the tedious process of the traditional design to repeatedly modify and validate with the aid of EDA tools, and shows less design redundancy, providing a reference for PDN high frequency circuit design.