基于同伦(homotopy)关系的ARMAX模型的辨识算法

论述了采用同伦 (hom otopy)法辨识 ARMAX模型的问题 .在 ARMAX模型的辨识中 ,由于滑动平均(MA)噪音模型部分的存在 ,用以估计 ARMAX模型参数的误差函数不再是单峰的 ,因而常用的基于最优化的估计算法有被陷入局部最小点的危险 .本文提出一种基于同伦关系的算法来解决这个问题 ,其基本想法是建立 ARMAX模型参数的估计和 ARX模型参数的估计两者之间的同伦关系 .由于用以估计 ARX模型参数的误差函数是单峰的 ,因而没有局部最小点问题 ,这就使得有可能利用它们之间的同伦关系来解决 ARMAX模型参数的估计中的局部最小点问题 .详细描述了基于同伦关系的新辨识算法 ,并利用蒙特卡罗 (Monte Car-lo)仿真试验说明了新算法的有效性

Homotopy approach to ARMAX model identification

This paper deals with the problem of identifying ARMAX model using homotopy approach. Because of the MA noise model part, the criterion function for ARMAX model identification is not always unimodal and an optimization based algorithm has a potential risk to be stuck at a local minimum. A homotopy approach is introduced to solve this problem. The idea is to construct a homotopy connection of the identification based on ARMAX model with the one based on ARX model.Monte Carlo simulations illustrate the effectiveness of the homotopy based method.