多视图聚类方法随着数据获取途径日益多样化成为研究热点,但大多数聚类方法低估了噪声和数据多结构互补性信息对聚类结果的影响,并且忽略了聚类结果对低秩张量优化过程的反向引导作用。为解决这些问题,提出了基于结构化张量学习的多视...多视图聚类方法随着数据获取途径日益多样化成为研究热点,但大多数聚类方法低估了噪声和数据多结构互补性信息对聚类结果的影响,并且忽略了聚类结果对低秩张量优化过程的反向引导作用。为解决这些问题,提出了基于结构化张量学习的多视图聚类(multi-view clustering based on structured tensor learning,MCSTL)。首先,对初始表示张量进行再次去噪使其更具准确性和鲁棒性;同时,互补地学习局部结构、全局结构和各视图间的高阶相关性,提高表示张量与原始数据本质簇结构的一致性;然后,从跨视图信息融合的亲和矩阵中学习到统一的特征矩阵,利用其隐含的聚类结构信息反向引导表示张量的优化过程;最后,对特征矩阵施加了正交约束,使其提供数据的软标签信息,并对模型进行直接聚类解释。实验表明,MCSTL在6种聚类评价指标上均表现优异,30个指标数据中有27个达到最优,从而充分验证了MCSTL的有效性和优越性。展开更多
Estimating the spatial distribution of coseismic slip is an ill-posed inverse problem, and solutions may be extremely oscillatory due to measurement errors without any constraints on the coseismic slip distribution. I...Estimating the spatial distribution of coseismic slip is an ill-posed inverse problem, and solutions may be extremely oscillatory due to measurement errors without any constraints on the coseismic slip distribution. In order to obtain stable solution for coseismic slip inversion, regularization method with smoothness-constrained was imposed. Trade-off parameter in regularized inversion, which balances the minimization of the data misfit and model roughness, should be a critical procedure to achieve both resolution and stability. Then, the active constraint balancing approach is adopted, in which the trade-off parameter is regarded as a spatial variable at each model parameter and automatically determined via the model resolution matrix and the spread function. Numerical experiments for a synthetical model indicate that regularized inversion using active constraint balancing approach can provides stable inversion results and have low sensitivity to the knowledge of the exact character of the Gaussian noise. Regularized inversion combined with active constraint balancing approach is conducted on the 2005 Nias earthquake. The released moment based on the estimated coseismic slip distribution is 9.91×1021 N·m, which is equivalent to a moment magnitude of 8.6 and almost identical to the value determined by USGS. The inversion results for synthetic coseismic uniform-slip model and the 2005 earthquake show that smoothness-constrained regularized inversion method combined with active constraint balancing approach is effective, and can be reasonable to reconstruct coseismic slip distribution on fault.展开更多
This paper addresses the problem of fault detection（FD） for networked systems with access constraints and packet dropouts.Two independent Markov chains are used to describe the sequences of channels which are availa...This paper addresses the problem of fault detection（FD） for networked systems with access constraints and packet dropouts.Two independent Markov chains are used to describe the sequences of channels which are available for communication at an instant and the packet dropout process,respectively.Performance indexes H∞ and H_ are introduced to describe the robustness of residual against external disturbances and sensitivity of residual to faults,respectively.By using a mode-dependent fault detection filter（FDF） as residual generator,the addressed FD problem is converted into an auxiliary filter design problem with the above index constraints.A sufficient condition for the existence of the FDF is derived in terms of certain linear matrix inequalities（LMIs）.When these LMIs are feasible,the explicit expression of the desired FDF can also be characterized.A numerical example is exploited to show the usefulness of the proposed results.展开更多
Based on bounded network-induced time-delay, the networked control system is modeled as a linear time-variant singular system. Using the Lyapunov theory and the linear matrix inequality approach, the criteria for dela...Based on bounded network-induced time-delay, the networked control system is modeled as a linear time-variant singular system. Using the Lyapunov theory and the linear matrix inequality approach, the criteria for delay-independent stability and delay-dependent stability of singular networked control systems are derived and transformed to a feasibility problem of linear matrix inequality formulation, which can be solved by the Matlab LMI toolbox, and the feasible solutions provide the maximum allowable delay bound that makes the system stable. A numerical example is provided, which shows that the analysis method is valid and the stability criteria are feasible.展开更多
冷热电联产(combined cooling, heating and power,CCHP)系统与微电网的结合有利于促进消纳可再生能源,为了提升CCHP型微电网的经济性、环保性和稳定性,提出了两阶段优化调度模型。离线优化阶段基于需求侧响应策略,建立了基于归一化法...冷热电联产(combined cooling, heating and power,CCHP)系统与微电网的结合有利于促进消纳可再生能源,为了提升CCHP型微电网的经济性、环保性和稳定性,提出了两阶段优化调度模型。离线优化阶段基于需求侧响应策略,建立了基于归一化法向约束法的多目标规划模型,并用熵权-TOPSIS法筛选最优结果。在线优化阶段建立了基于动态矩阵控制算法的有限时域优化模型,对离线优化结果进行跟踪优化和反馈校正,以降低不确定性因素的影响。最后,设计对比方案进行分析,验证了所提优化模型的有效性。展开更多
自适应波束形成随着数字信号处理技术的不断发展,已广泛应用于雷达、语音、医疗等领域。然而,当阵列发生扰动时,将会导致干扰偏离零陷位置,甚至会导致算法完全失效。为了解决现有波束形成算法在发生导向矢量失配和干扰位置扰动时波束形...自适应波束形成随着数字信号处理技术的不断发展,已广泛应用于雷达、语音、医疗等领域。然而,当阵列发生扰动时,将会导致干扰偏离零陷位置,甚至会导致算法完全失效。为了解决现有波束形成算法在发生导向矢量失配和干扰位置扰动时波束形成器性能急剧下降的问题,本文提出了一种导向矢量失配条件下多约束鲁棒波束形成算法。本文参照实际情况引入更多约束,增加了双边范数扰动约束以及二次相似性约束,允许了误差产生的范围。此外,本文确保感兴趣信号(Signal Of Interest,SOI)的到达方向(Direction Of Arrival,DOA)远离干扰导向矢量的所有线性组合的DOA区域,保证了最优导向矢量的DOA位于SOI的角扇形区域。首先,以波束形成器输出最大功率为目标,并结合实际环境下的约束条件,建立了最优导向矢量的数学模型。其次,利用定义的干扰范围重构协方差矩阵,以此来展宽零陷,提高系统的抗干扰性能。最后,先用内点法求得替代变量的解,以此求解针对导向矢量的二次不等式约束问题;随后在约束模型中代入替代变量,用交替方向乘子法迭代求解导向矢量,在每一次的迭代中都会得到显示解。同时,本文还对算法的时间复杂度和收敛性进行了分析。实验结果显示,相较于传统的波束形成算法,所提方法加宽了干扰处零陷,使得波束形成器的抗干扰性能得到了一定的提高,且能够很好地校正失配导向矢量。展开更多
文摘多视图聚类方法随着数据获取途径日益多样化成为研究热点,但大多数聚类方法低估了噪声和数据多结构互补性信息对聚类结果的影响,并且忽略了聚类结果对低秩张量优化过程的反向引导作用。为解决这些问题,提出了基于结构化张量学习的多视图聚类(multi-view clustering based on structured tensor learning,MCSTL)。首先,对初始表示张量进行再次去噪使其更具准确性和鲁棒性;同时,互补地学习局部结构、全局结构和各视图间的高阶相关性,提高表示张量与原始数据本质簇结构的一致性;然后,从跨视图信息融合的亲和矩阵中学习到统一的特征矩阵,利用其隐含的聚类结构信息反向引导表示张量的优化过程;最后,对特征矩阵施加了正交约束,使其提供数据的软标签信息,并对模型进行直接聚类解释。实验表明,MCSTL在6种聚类评价指标上均表现优异,30个指标数据中有27个达到最优,从而充分验证了MCSTL的有效性和优越性。
基金Projects(41604111,41541036) supported by the National Natural Science Foundation of China
文摘Estimating the spatial distribution of coseismic slip is an ill-posed inverse problem, and solutions may be extremely oscillatory due to measurement errors without any constraints on the coseismic slip distribution. In order to obtain stable solution for coseismic slip inversion, regularization method with smoothness-constrained was imposed. Trade-off parameter in regularized inversion, which balances the minimization of the data misfit and model roughness, should be a critical procedure to achieve both resolution and stability. Then, the active constraint balancing approach is adopted, in which the trade-off parameter is regarded as a spatial variable at each model parameter and automatically determined via the model resolution matrix and the spread function. Numerical experiments for a synthetical model indicate that regularized inversion using active constraint balancing approach can provides stable inversion results and have low sensitivity to the knowledge of the exact character of the Gaussian noise. Regularized inversion combined with active constraint balancing approach is conducted on the 2005 Nias earthquake. The released moment based on the estimated coseismic slip distribution is 9.91×1021 N·m, which is equivalent to a moment magnitude of 8.6 and almost identical to the value determined by USGS. The inversion results for synthetic coseismic uniform-slip model and the 2005 earthquake show that smoothness-constrained regularized inversion method combined with active constraint balancing approach is effective, and can be reasonable to reconstruct coseismic slip distribution on fault.
基金supported by the National Natural Science Foundation of China (6057408860874053)
文摘This paper addresses the problem of fault detection（FD） for networked systems with access constraints and packet dropouts.Two independent Markov chains are used to describe the sequences of channels which are available for communication at an instant and the packet dropout process,respectively.Performance indexes H∞ and H_ are introduced to describe the robustness of residual against external disturbances and sensitivity of residual to faults,respectively.By using a mode-dependent fault detection filter（FDF） as residual generator,the addressed FD problem is converted into an auxiliary filter design problem with the above index constraints.A sufficient condition for the existence of the FDF is derived in terms of certain linear matrix inequalities（LMIs）.When these LMIs are feasible,the explicit expression of the desired FDF can also be characterized.A numerical example is exploited to show the usefulness of the proposed results.
基金the National Natural Science Foundation of China (60574011)the National Natural Science Foundation of Liaoning Province (2050770).
文摘Based on bounded network-induced time-delay, the networked control system is modeled as a linear time-variant singular system. Using the Lyapunov theory and the linear matrix inequality approach, the criteria for delay-independent stability and delay-dependent stability of singular networked control systems are derived and transformed to a feasibility problem of linear matrix inequality formulation, which can be solved by the Matlab LMI toolbox, and the feasible solutions provide the maximum allowable delay bound that makes the system stable. A numerical example is provided, which shows that the analysis method is valid and the stability criteria are feasible.
文摘冷热电联产(combined cooling, heating and power,CCHP)系统与微电网的结合有利于促进消纳可再生能源,为了提升CCHP型微电网的经济性、环保性和稳定性,提出了两阶段优化调度模型。离线优化阶段基于需求侧响应策略,建立了基于归一化法向约束法的多目标规划模型,并用熵权-TOPSIS法筛选最优结果。在线优化阶段建立了基于动态矩阵控制算法的有限时域优化模型,对离线优化结果进行跟踪优化和反馈校正,以降低不确定性因素的影响。最后,设计对比方案进行分析,验证了所提优化模型的有效性。
文摘自适应波束形成随着数字信号处理技术的不断发展,已广泛应用于雷达、语音、医疗等领域。然而,当阵列发生扰动时,将会导致干扰偏离零陷位置,甚至会导致算法完全失效。为了解决现有波束形成算法在发生导向矢量失配和干扰位置扰动时波束形成器性能急剧下降的问题,本文提出了一种导向矢量失配条件下多约束鲁棒波束形成算法。本文参照实际情况引入更多约束,增加了双边范数扰动约束以及二次相似性约束,允许了误差产生的范围。此外,本文确保感兴趣信号(Signal Of Interest,SOI)的到达方向(Direction Of Arrival,DOA)远离干扰导向矢量的所有线性组合的DOA区域,保证了最优导向矢量的DOA位于SOI的角扇形区域。首先,以波束形成器输出最大功率为目标,并结合实际环境下的约束条件,建立了最优导向矢量的数学模型。其次,利用定义的干扰范围重构协方差矩阵,以此来展宽零陷,提高系统的抗干扰性能。最后,先用内点法求得替代变量的解,以此求解针对导向矢量的二次不等式约束问题;随后在约束模型中代入替代变量,用交替方向乘子法迭代求解导向矢量,在每一次的迭代中都会得到显示解。同时,本文还对算法的时间复杂度和收敛性进行了分析。实验结果显示,相较于传统的波束形成算法,所提方法加宽了干扰处零陷,使得波束形成器的抗干扰性能得到了一定的提高,且能够很好地校正失配导向矢量。
