基于改进遗传算法的模拟电路参数自动化设计  

Automated design of analog circuit based on improved genetic algorithm

作　　者：李红磊[1] 刘家学[1] 

机构地区：[1]中国民航大学航空自动化学院,天津300300

出　　处：《电子设计工程》2015年第17期147-150,共4页Electronic Design Engineering

基　　金：国家科技支撑计划(2012BAC20B03);国家自然科学基金重点项目(60832011);天津市科技攻关计划重点项目(06YFGZGX00700)

摘　　要：模拟电路的元件参数很多,而且设计时需要兼顾多个性能要求,导致计算调整困难。针对此问题,提出一种以遗传算法为基础的参数自动设计方法。首先,以性能达标程度的加权和作为适应度评价函数,为不同结构、不同性能要求的模拟电路建立了形式较为统一的数学评价模型;其次,采用子代竞争择优、精英超突变等策略改进遗传算法,提高了算法的收敛速度和局部搜索能力,并对遗传算法求出的元件参数做了可行化处理,使其满足实际制作的需要;最后,利用该方法设计制作中心频率为115 MHz的电压控制振荡器,验证方法有效性。Analog circuit has a lot of parameters which should consider the multiple performance objectives when were designed, this cause the difficulty in calculation and adjustment. In order to solve the problem, a method for designing the parameters of structure known circuits was presented based on GA （ genetic algorithm）.Firstly, using weighted summation of performance as the fitness evaluation function, for different analog circuits and different performance requirements of analog circuit, a unified mathematical evaluation model is established. Secondly, some strategies, such as the offspring competitive, elitist and hyper-mutation strategy, were used to improve the GA＇ s convergence speed and local searching ability. To meet the needs of actual production, some feasible treatments on parameters calculated by GA have been done. Based on the above method, a voltage-controlled oscillator （the center frequency is 115 MHz）is designed, to confirm the effective of the program.

关 键 词：模拟电路设计 遗传算法 性能达标程度 超突变 可行化 

分 类 号：TN7[电子电信—电路与系统]