基于遗传算法的亚100nm SOI MOSFET模型参数提取  

Genetic-Algorithm-Based Model Parameter Extraction for Sub-100nm SOI MOSFET

作　　者：李尊朝[1] 张瑞智[1] 张效娟[2] 林尧[1] 

机构地区：[1]西安交通大学电子与信息工程学院,陕西西安710049 [2]青海师范大学计算机系,青海西宁810008

出　　处：《电子学报》2007年第11期2033-2037,共5页Acta Electronica Sinica

基　　金：国家自然科学基金(No.NSFC60472003);国家973重点基础研究发展计划(No.2005CB321701)

摘　　要：为了简化亚100nm SOI MOSFET BSIMSOI4的模型参数提取过程,实现全局优化,使用了遗传算法技术,并提出了保留多个最优的自适应遗传算法.该算法通过保留最优个体的多个拷贝,对适应度高和适应度低的个体分别进行诱导变异和动态变异,在进化起始阶段和终止阶段分别执行随机交叉和诱导交叉,既具有全局优化特性,又加速了局部搜索过程,提高了最终解的质量.不同种群数和进化代数条件下的参数提取实例表明,该算法提取精度高、速度快,全局优化稳定性好;适当增加种群数,有利于加速算法的全局收敛过程.Genetic algorithm is used in BSIMSOI4 model parameter extraction for sub-100nm SOI MOSFETs to simplify extraction process and optimize parameters globally.An extraction algorithm called adaptive genetic algorithm maintaining multi-optimum is proposed.In the new algorithm,multiple copies of the optimum chromosome in each generation are kept,induced and dynamic mutations are carried out on chromosomes with larger and smaller fitness,respectively,and random and induced crossovers are executed in the early and late generations,respectively.The global optimization is maintained,the local searching is speeded up and the quality of the final solution is improved.Extraction examples with different population sizes and evolutionary steps show that the new algorithm has the advantages of higher accuracy,faster convergence,and reliable global optimization and that global convergence could be speeded up by increasing population sizes properly.

关 键 词：SOI 参数提取 全局优化 模型 

分 类 号：TN432[电子电信—微电子学与固体电子学]