摘要
基于自适应遗传算法,提出一种多项式模型结构与参数的一体化辨识方法.针对组合非线性系统,首先将选定的候选项原始序列与输出序列进行相关度评估,根据其大小排列进行遗传算法染色体结构的自适应编码;在迭代辨识充分后,再次计算候选项贡献序列与由该项造成的模型损失序列间的相关度,剔除相关度较小的项,调整模型结构;如此循环迭代,在完成参数辨识的同时最终确认模型结构.仿真实例验证了算法的有效性.
A method of structure validation and parameter estimation based on adaptive genetic algorithm is proposed.For an assembled polynomial nonlinear system,the correlation coefficient between the selected candidate item’s initial sequence and its output sequence is calculated,and the chromosome’s structure of genetic algorithm is coded adaptively according to the correlation coefficient array.After that,the correlation coefficient between the candidate item’s contribution sequence and the lost sequence that caused by this item is calculated,the items with smaller correlation coefficients are eliminated,and the model’s structure is regulated.Finally the model’s structure and parameters can be obtained by repeating the above steps.Several simulation results show the effectiveness of the method.
出处
《控制与决策》
EI
CSCD
北大核心
2011年第5期761-767,共7页
Control and Decision
关键词
非线性系统辨识
自适应遗传算法
相关系数
结构辨识
多项式模型
智能计算
nonlinear system identification
adaptive genetic algorithm
correlation coefficient
structure validation
polynomial model
intelligent computation
