摘要
为了准确地评价多变量控制系统的性能,并简化其评价过程的计算复杂度,提出了基于主元分析(PCA)与最小二乘支持向量机(LSSVM)相结合的多变量控制系统性能评价方法。该方法将原始自变量数据通过PCA方法进行降维处理,利用二次损失函数取代支持向量机中的不敏感损失函数,将不等式约束变为等式约束,从而将二次规划问题转变为线性方程组的求解,并对LSSVM的参数选取做了改进。该方法在性能评价过程中不需要求解系统关联矩阵,简化了求解的复杂性。仿真实例验证了PCA-LSSVM性能评价方法更能反映控制系统真实性能。
In order to evaluate accurately the performance of multi-variable control system and simplify the computational complexity of the evaluation process, the performance assessment method based on the combination of principal component analysis （ PCA） and least square support vector machine （LSSVM） for multi-variable control system is proposed. The dimension reduction of the original independent variables is conducted by PCA method; the insensitive loss function in SVM is replaced by quadratic loss function, and the inequality constraints is replaced by equality constraints, thus the quadratic programming is converted into the solution of linear equations; in addition, the selection of parameters of ISSVM is improved. The solving of system associated matrix is not necessary with this method for performance assessment, and the solving complexity is simplified. The practical example of simulation verifies that the PCA-LSSVM method better reflects the real performance of the control system.
出处
《自动化仪表》
CAS
北大核心
2014年第1期10-14,共5页
Process Automation Instrumentation
基金
辽宁省科技攻关基金资助项目(编号:2011216011)
关键词
最小二乘支持向量机
多变量控制系统
主元分析
关联矩阵
性能指标
性能评价
Least square support vector machine Muhi-variable control system Principal component analysis Associated matrixPerformance indexes Performance evaluation
