全文下载:
201906015.pdf
文章编号: 1672-6987(2019)06-0106-05; DOI: 10.16351/j.1672-6987.2019.06.015
姜晓坤a, 籍艳a*, 万立娟b
(青岛科技大学 a.自动化与电子工程学院; b.数理学院,山东 青岛 266061)
摘要: 提出了一种输出误差滑动平均(OEMA)系统的两阶段增广随机梯度算法。利用辅助模型思想处理未知变量,并用随机梯度法识别系统模型参数。应用分解的技术将OEMA系统分解为两个低维数的子系统,并分别识别每个子系统。由于子系统中协方差矩阵的维数降低,因此减少了计算量,从而提高了算法的计算效率。仿真结果表明该算法是有效的。
关键词: 辅助模型; 随机梯度; 参数估计; 分层识别; 计算量
中图分类号: TP 273文献标志码: A
引用格式: 姜晓坤, 籍艳, 万立娟. 输出误差滑动平均系统的递阶增广随机梯度算法\[J\]. 青岛科技大学学报(自然科学版), 2019, 40(6): 106-110.
JIANG Xiaokun, JI Yan, WAN Lijuan. Hierarchical extended stochastic gradient algorithm for output error moving average systems\[J\]. Journal of Qingdao University of Science and Technology(Natural Science Edition), 2019, 40(6): 106-110.
Hierarchical Extended Stochastic Gradient Algorithm for
Output Error Moving Average SystemsJIANG Xiaokuna, JI Yana, WAN Lijuanb
(a.College of Automation and Electronic Engineering;b.College of Mathematics and Physics,
Qingdao University of Science and Technology,Qingdao 266061,China)
Abstract: This paper presents a two-stage extended stochastic gradient algorithm for output error moving average (OEMA) systems.The basic idea is to use the auxiliary model idea to deal with unknown variables,and to use the stochastic gradient method to identify system model parameters.The OEMA system has been decomposed into two subsystems and identify each subsystem,respectively.The dimension of the covariance matrix involved in the subsystem becomes smaller,the calculation amount is reduced,and the proposed algorithm has higher computational efficiency.The simulation results indicate that the proposed algorithm is effective.
Key words: auxiliary model; stochastic gradient; parameter estimation; hierarchical identification;the load of calculation
收稿日期: 2018-10-29
基金项目: 山东省自然科学基金项目(ZR201702170236).
作者简介: 姜晓坤(1994—),男,硕士研究生.*通信联系人.