基于线性规划问题递推算法的L_∞参数辨识  

L∞ parameter identification based on linear programming recursive algorithm

作　　者：肖建华[1] 赵明旺[2] 耿修堂[1] 陈智华[1] 

机构地区：[1]华中科技大学图像信息处理与智能控制教育部重点实验室,湖北武汉430074 [2]武汉科技大学信息科学与工程学院,湖北武汉430081

出　　处：《系统工程与电子技术》2007年第8期1339-1341,共3页Systems Engineering and Electronics

摘　　要：针对传统的最小二乘辨识算法要求误差遵循零均值、同方差的正态分布等不足,提出了L∞参数辨识算法。首先将L∞参数辨识问题转化成增加约束条件的线性规划问题最优解的求解问题;然后为了删除辨识中的冗余数据并用于在线辨识,给出了基于线性规划问题递推算法的L∞参数辨识算法。最后给出了仿真,结果验证了算法的有效性。To overcome the shortages of the traditional least squares algorithms that require error follow zero mean valued and homo-skedastcity normal distribution, a L∞ identification method is presented. Firstly. the L∞ identification is transferred into the optimal solution for the increase of constraints to the linear programming. Then in order to delete the redundant data of identification and the application of on-line identification, the L∞ parameter identification based on linear programming recursive algorithm is given. Finally. a numerical example of simulation is given and the computing results show the effectiveness of the proposed algorithm.

关 键 词：线性规划 L∞辨识 递推辨识 

分 类 号：N945.14[自然科学总论—系统科学]