基于改进斑马算法的GaN HEMT 混合小信号建模  

GaN HEMT hybrid small-signal modeling based on improved Zebra algorithm

作　　者：李畅 王军[1] LI Chang;WANG Jun(College of Information Engineering,Southwest University of Science and Technology,Mianyang 621010,Sichuan Province,China)

机构地区：[1]西南科技大学信息工程学院,四川绵阳621010

出　　处：《电子元件与材料》2025年第1期49-56,共8页Electronic Components And Materials

基　　金：国家自然科学基金(69901003);四川省教育厅自然科学基金(18ZA0502)。

摘　　要：为了提高半导体器件小信号建模精度并解决优化算法易陷入局部最优解的问题,提出了一种基于改进斑马优化算法(Improved Zebra Optimization Algorithm,IZOA)的氮化镓高电子迁移率晶体管(Gallium Nitride High Electron Mobility Transistor,GaN HEMT)混合小信号建模方法。采用数学修正法和直接提取法提取小信号参数,建立初步模型,再使用改进的斑马优化算法进一步提高建模的精度。对斑马优化算法(Zebra Optimization Algorithm,ZOA)的改进主要集中在三个方面:采用混沌映射提高初始种群多样性;使用反向学习策略扩大搜索范围;使用动态概率值替代固定值平衡搜索与收敛能力。实验结果表明,IZOA将直接提取法的平均误差从3.47%降至0.19%,相比灰狼优化(Grey Wolf Optimizer,GWO)算法(平均误差0.95%)降低0.76%,较标准ZOA(平均误差0.52%)降低0.33%,验证了算法的有效性和准确性。To enhance small-signal modeling precision of the semiconductor device and avoid the local optimum of the optimization algorithm,a hybrid small-signal modeling method for Gallium Nitride High Electron Mobility Transistor(GaN HEMT)with Improved Zebra Optimization Algorithm(IZOA)was proposed.Small-signal parameters were extracted by the mathematical correction method and the direct extraction method to establish an initial model.The improved zebra optimization algorithm was apply to further boost modeling accuracy.The Zebra Optimization Algorithm(ZOA)improvements focused on three aspects:adopt chaotic mapping for initial population diversity;apply opposition-based learning strategy to enlarge search range;employ dynamic probability values rather than fixed to balance search and convergence.The experimental results show that,the average error of direct extraction method can be decreased from 3.47%to 0.19%by IZOA.Compared with the Grey Wolf Optimizer(GWO)algorithm(average error 0.95%),it is reduced by 0.76%,and it is 0.33%lower than that of the ZOA(average error 0.52%).Thus,the effectiveness and accuracy of the algorithm were verified.

关 键 词：GaN HEMT 小信号模型 斑马优化算法 参数提取方法 改进算法 

分 类 号：TN386[电子电信—物理电子学]