用遗传算法优化设计分数阶微分器  

Genetic Algorithm for Optimal Design of Fractional Order Differentiator

作　　者：张力支[1] 周激流[2] 郎方年[3] 杨柱中[1] 雷印杰[1] 

机构地区：[1]四川大学电子信息学院,四川成都610064 [2]四川大学计算机学院软件学院,四川成都610064 [3]重庆大学计算机学院,重庆400044

出　　处：《四川大学学报（工程科学版）》2008年第1期158-162,共5页Journal of Sichuan University (Engineering Science Edition)

基　　金：国家自然科学基金资助项目(60572033)

摘　　要：为了进一步提高分数阶微分器的精度,将遗传算法引入到滤波设计中,获得了性能更加优越的分数阶微分器。首先分析了分数阶微分器的特点和设计方法,通过对相应IIR滤波器的系数进行优化,得到了分数阶微分器的全局最优个体,避免了传统方法设计分数阶微分器中高频出现的误差。仿真结果表明,由该算法设计的分数阶微分器明显优于其它方法,能有效的逼近理想分数阶微分器的幅频响应。;In order to improve the accuracy of fractional order differentiator, a hybrid genetic algorithm was applied to design fractional order differentiator. The characters and design method of fractional order differentiator was analyzed, and through optimizing coefficients of corresponding IIR filter, the global optimum chromosome was obtained and high frequency deviation of fractional order differentiator designed by traditional methods was avoided. Simulation experiments demonstrated that the fractional order differentiator performance obtained in this paper outperforms to others and its magnitude response approaches to the ideal fractional order differentiator.

关 键 词：分数阶微分器 遗传算法 分数阶微积分 分数阶控制系统 变异系数 

分 类 号：TN713[电子电信—电路与系统] TP18[自动化与计算机技术—控制理论与控制工程]