论文部分内容阅读
提出一种基于改进遗传算法和递推最小二乘的非线性模糊辨识新算法。该辨识方法包含结构辨识辨出和参数辨识,结构辨识即输入空间的模糊划分,采用具有自适应性的广义高斯隶属函数;参数辨识包含前提参数和结论参数,用基于动态比例变换的改进遗传算法优化高斯函数的前提参数,用递推最小二乘辨识模糊模型的结论参数。最后通过著名的Box-Jenkins煤气炉数据仿真(仿真环境:MATLAB 6.5,计算机主频2.4 GHz,内存512 MB),并根据输入变量个数和模糊规则数,得到均方误差以证明本文方法的辨识精度,将该文辨识方法与其他方法进行比较,验证了该方法辨识精度更高。
A new non-linear fuzzy identification algorithm based on improved genetic algorithm and recursive least squares is proposed. The identification method includes structure identification identification and parameter identification, structure identification is fuzzy partitioning of input space, adaptive generalized Gaussian membership function is adopted, parameter identification includes precondition and conclusion parameters, and dynamic genetic algorithm based on dynamic proportional transformation Optimize the precondition parameters of Gaussian function, and use recursive least squares to identify the conclusion parameters of the fuzzy model. Finally, through the famous Box-Jenkins gas stove data simulation (simulation environment: MATLAB 6.5, computer frequency 2.4 GHz, memory 512 MB), and according to the number of input variables and the number of fuzzy rules, the mean square error is obtained to prove the identification of this method The accuracy of this method is compared with that of other methods. The result shows that this method has better identification accuracy.