摘要
针对传统方法存在响应时间较长、绝对误差较大等问题,提出了一种基于动态主元分析法的不协调目标信息系统多变量控制方法。根据主元分析方法引入折息因子,建立多个参数可调的动态主元分析算法,通过该算法计算各个变量的特征值以及贡献率,确定不协调目标信息系统的主元数量,构建目标信息系统多变量控制模型。采用粒子概率假设密度滤波器对目标信息系统多变量控制模型进行求解,在信息增益最大化的准则下获取目标信息系统的最优控制方案。实验结果表明,所提方法能够高效、准确地对不协调目标信息系统的多变量进行控制,达到预期的控制效果。
Due to long response time and large absolute error in traditional methods, this paper proposed a multi-variable control method in uncoordinated objective information system based on dynamic principal component analysis. Based on the principal component analysis, discount factor was introduced to establish some dynamic principal component analysis algorithms with adjustable parameter. The eigenvalues and contribution rates of each variable were calculated by this algorithm, and then the number of principal components in uncoordinated objective information system was determined. Moreover, the multivariable control model of objective information system was constructed. The particle probability was used to assume density filter, so that multivariable control model of objective information system was solved. Based on the criterion of maximizing information gain, the optimal control scheme of objective information system was obtained. Simulation results show that the proposed method can effectively and accurately control the multivariables of uncoordinated objective information system. The expected control effect is achieved.
作者
黄丽娟
邹文君
吴凡
HUANG Li-juan;ZOU Wen-jun;WU Fan(Institute of Technology East China Jiao Tong University,Nanchang Jiangxi 330100,China)
出处
《计算机仿真》
北大核心
2019年第7期187-190,212,共5页
Computer Simulation
基金
2017年度江西省教育厅科学技术研究项目(GJJ171485)
2018年度江西省高校人文社会科学研究项目(TQ18106)
关键词
不协调
目标信息系统
多变量控制方法
Uncoordinated
Objective information system
Multivariate control method
