应用混合遗传算法的超磁致伸缩致动器磁滞模型的参数辨识

PARAMETER IDENTIFICATION OF HYSTERETIC MODEL FOR GINAT MAGNETOSTRICTIVE ACTUATOR USING HYBRID GENETIC ALGORITHM

针对遗传算法爬山能力差的弱点,该文把信赖域算法作为一个与选择、交叉和变异平行的算子,嵌入到遗传算法中,得到一种混合计算智能算法。该方法兼顾了遗传算法和信赖域算法的长处,既有较快的收敛速度,又能以非常大的概率求得最优解。应用该混合遗传算法对超磁致伸缩致动器的磁滞非线性动态模型进行参数辨识,仿真和实验研究表明,该算法能有效地辨识非线性系统的非线性参数,并具有一定的抗噪声能力。

Aiming at the weak capacity of climbing hill of genetic algorithm, a hybrid intelligent algorithm is established by setting the trust region algorithm in the genetic algorithm. In the proposed hybrid genetic algorithm, the trust region algorithm is taken as a genetic operator which parallels to the selection, crossover and mutation operators. The hybrid algorithm is paid attention to both the advantages of the trust region algorithm and the genetic algorithm. It not only has a rather high convergence speed, but also can locate the best solution with a rather large probability. This paper applies the hybrid genetic algorithm to identify parameters of dynamic model with hysteretic nonlinearity for giant magnetostrictive actuator. The simulation and experimental results show that the hybrid algorithm can efficiently identify nonlinear parameters of nonlinear systems even with system noise.