论文部分内容阅读
本文研究了基于遗传算法的系统辨识的可靠性。首先,分析了基于响应曲线拟合的系统辨识方法及各种模型的拟合精度对模型参数摄动的敏感度。然后,对比分析了几种遗传算法应用于此类系统辨识方法的可靠性及效率,并给出了提高系统辨识效率及可靠性的改进遗传策略。实验结果表明基于改进遗传算法的系统辨识方法的可靠性乃有效性。
This paper studies the reliability of system identification based on genetic algorithm. First of all, the sensitivity of model parameters perturbation to model parameters based on response curve fitting method and the fitting accuracy of various models are analyzed. Then, the reliability and efficiency of several genetic algorithms applied to such system identification are compared and analyzed, and the improved genetic strategy to improve the system identification efficiency and reliability is given. Experimental results show that the reliability of the system identification method based on improved genetic algorithm is effective.