基于改进PSO-SMO的模型参数辨识及应用

摘要

目前,大多数智能辨识方法只考虑辨识系统的输出值,不能辨识模型的参数.对此,研究了基于线性核函数支持向量机的辨识算法,对模型参数和输出同时进行辨识.在此基础上,采用改进的PSO-SMO算法以提高辨识速度和精度.将该方法用于ARX模型和长期预测模型的参数辨识中,结果表明,该算法比其他算法具有更高的准确性.%At present, the most intelligent identification methods can only identify die output of a system, instead of the model parameters. For this problem, this paper studied the identification algorithm of SVM based on linear kernel function. In the concrete realization, it used the improved PSO-SMO algorithm to improve identification speed and precision. Using this method to identify the parameters of ARX model and long-term prediction model, the simulation results show that this algorithm has a higher accuracy than that of other method.