基于MPI的全局并行遗传算法的SOI MOS器件模型参数提取

SOI MOSFET Model Parameter Extraction Based on Hybrid Parallel Genetic Algorithm

提出了将全局并行遗传算法应用于模型参数提取,并应用于标准的1.2μm CMOS/SOI工艺的SOI MOS-FET器件,一次性提取BSIMSOI3模型主要的42个直流参数.实验结果表明,该方法不依赖参数初始值、精度高、效率高,降低了SOI模型参数提取工作的难度,具有很强的通用性,易于移植优化及数据拟合.

A MPI - based Global Parallel Genetic Algorithm method is put forward for extracting model parameters by global optimization strategy. Simulation result is compared with measurement data of 1.2μm CMOS/SOI. The simulation result fits the t data very well. The method is very simple and no plenty experience about parameter extraction. It is convenient to apply it to other complex parameter extraction fields.