基于连续域蚁群改进算法的T-S模型辨识  

Identification of T-S fuzzy model based on improved ACS in continuous space

作　　者：张浩[1] 彭晓华[2] 孟宇[1] 

机构地区：[1]辽宁工程技术大学电气与控制工程学院,辽宁葫芦岛125105 [2]辽宁工程技术大学理学院,辽宁阜新123000

出　　处：《黑龙江科技学院学报》2007年第1期74-78,共5页Journal of Heilongjiang Institute of Science and Technology

摘　　要：针对一些带复杂参数辨识性能指标的模糊建模无法采用经典辨识方法解决的情况,提出了采用改进连续域蚁群算法进行T-S模型辨识的方法。该方法利用蚁群算法无需具体数学模型的特点,解决了模糊模型辨识中辨识指标表达式过于复杂甚至非解析所带来的难题,从而实现了T-S模型前提部和结论部参数的一次性辨识。仿真研究表明,利用文中辨识方法得到的模糊模型能够很好的逼近真实系统,并且对初始条件和输入信号不敏感,具有鲁棒性。Some fuzzy identification problem can not be solved by traditional methods, because of their complex expression of performance index. A new identification method of fuzzy model using improved ACS in continuous space is proposed in this paper. ACS can optimize systems without their mathematic models, so the method proposed in this paper uses it to solve the problem-the expression of performance index for identification is non-analytic. Based on this, the premise part and consequence part in the T-S model can be identified at one time. The simulated experiment shows that the fuzzy model based on the identification method proposed in this paper can well trace the real system. In addition, it is insensitive to the initial condition and the input signal.

关 键 词：T-S模型 连续域蚁群算法 系统辨识 

分 类 号：TP273.4[自动化与计算机技术—检测技术与自动化装置]