Stability analysis of positive switched systems via joint linear copositive Lyapunov functions

In this paper, we study the asymptotic stability of continuous-time positive switched linear systems for the case when each subsystem is only stable. By using the so-called “joint linear copositive Lyapunov function” (JLCLF) generalizing the common linear copositive Lyapunov function, we show that the system remains asymptotically stable under appropriate switching if it has a JLCLF. Then, the main result is extended to positive switched linear systems with time delay.     

Classification Method of Coal and Gangue Using Terahertz Time-Domain Spectroscopy,Cluster Analysis and Principal Component Analysis

Journal of Applied Spectroscopy - The process of coal mining generates high amounts of coal gangue. Accordingly, coal–gangue separation is a key problem limiting coal production and quality....     

Continuous output integral sliding mode control for switched linear systems

A continuous output integral sliding mode robustification methodology for switched uncertain linear time invariant systems with state-dependent location transitions and dwell time is presented. The robustifying methodology is based on the adjustment of the super-twisting algorithm gains to assure the convergence time and the attenuation of the chattering. The use of the adjusted STA allows to reconstruct the states theoretically exactly before half of the dwell time without the usage of filters, via a continuous cascade observer. Moreover, it allows to generate a continuous control signal that is turn on after the observer has converged, guaranteeing theoretically exact compensation of the matched uncertainties/perturbations before the dwell time.     

铁水含硅量预测系统中数据预处理技术的研究

通过使用自适应预报模型与时差方法相结合的方法对铁水含硅量进行预测,模型所需的工艺参数的原始数据经采集后需使用OPC技术标准进行传输以存入数据库系统中。硅预测程序通过对数据库查询得到所需的操作数据,由于生产环境对数据的干扰,需对其作均值滤波和插值运算的一次处理,然后进行求平均值、梯度和标准差统计学运算的二次处理,之后运用模糊理论将各个工艺参数归一化以提取参数特征值,作为铁水含硅量预测模型的输入参数,在预测过程中使用多元线性回归不断修正模型中的权重系数,以提高数学模型预测的准确率。     

New error bounds for the linear complementarity problem with an SB-matrix

Error bounds for SB-matrices linear complementarity problems are given in the paper (Dai et al., Numer Algorithms 61:121–139, 2012). In this paper, new error bounds for the linear complementarity problem when the matrix involved is an SB-matrix are presented and some sufficient conditions that new bounds are sharper than those of the previous paper under certain assumptions are provided. New perturbation bounds of SB-matrices linear complementarity problems are also considered.     

基于CARS-MIV-SVR的库尔勒香梨可溶性固体含量预测方法

为了实现库尔勒香梨依据可溶性固体含量（SSC）分级定等和按质论价,推动采后处理向标准化、产业化方向健康发展,利用高光谱成像技术研究出了一种快速、有效、无损检测库尔勒香梨SSC的方法。以表面无损伤的157个库尔勒香梨作为研究样本,应用高光谱成像采集系统获取400~1 000 nm波长范围内高光谱图像并用ENVI5.3软件提取感兴趣区域（ROI）,获得高光谱数据。采用Kennard-Stone（KS）样本集划分方法将全部样本按照2∶1的比例划分为校正集（105）和预测集（52）。对比标准变量变换（SNV）、多元散射校正（MSC）、一阶导数（FD）和二阶导数（SD）等数据预处理方法对建模精度的影响,最终选用SNV方法对光谱曲线进行平滑去噪。该研究提出竞争性自适应重加权算法与平均影响值算法的组合算法（CARS-MIV）选择特征波长。在竞争性自适应重加权算法（CARS）方法中,建模样本由蒙特卡罗算法随机选择生成,变量回归系数会随之发生变化,因而回归系数的绝对值不能全面反映变量重要性,从而影响模型检测精度。为降低这种影响,应用平均影响值（MIV）算法对选出的自变量进行二次筛选,筛选出相关性较大的变量用以建模分析,并与CARS、连续投影算法（SPA）、蒙特卡罗无信息变量消除算法（MCUVE）等经典特征波长选择算法进行比较。最后分别以全波长（FS）光谱信息和四种特征波长选择方法得出的光谱信息作为输入矢量,应用支持向量回归（SVR）建立库尔勒香梨可溶性固体含量定量预测数学模型,以校正集相关系数（R_c）、校正集均方根误差（RMSEC）、预测集相关系数（R_p）和预测集均方根误差（RMSEP）四个参数来评估模型的预测精度。比较分析发现,CARS-MIV-SVR模型效果最佳,校正集相关系数（R_c）为0.985 94,预测集相关系数（R_p）达到0.946 31,校正集和预测集均方根误差分别为0.185 85和0.403 33。结果证明：CARS-MIV特征波长选择方法能够有效增强库尔勒香梨光谱数据特征波长选择的稳定性和精确性,提高模型的预测精度。利用高光谱技术结合CARS-MIV-SVR模型能够满足库尔勒香梨可溶性固体含量测定需求,实现库尔勒香梨的分级定等和按质论价。