广义次正定矩阵上的Oppenheim不等式

用矩阵分析的方法, 通过对广义次正定矩阵性质的进一步研究, 得到了更一般条件下的两个广义次正定矩阵的Hadamard乘积的行列式下界估计的Oppenheim不等式, 在适用范围和估计精度上都改进了已有的相应结果.     

基于两级神经网络的发酵过程多变量前馈解耦控制

针对具有时变、非线性、不确定性的多变量耦合生物发酵过程,提出了一种基于两级神经网络的多变量前馈解耦方法.一级神经网络利用可获得的过程信息拟和耦合通道的过程特性,实现耦合作用对被控量影响的估计;二级神经网络用来拟和控制通道的逆特性.通过两级网络的串联,消除系统间的耦合.实验结果表明,提出的解耦控制方法能适应生物发酵过程模型的不确定性和参数时变性,克服了前馈解耦方法依赖于过程模型和对模型参数的变化表现敏感的缺点.     

基于过程神经网络的短期负荷预测

针对目前常用方法在解决负荷预测问题时,结果往往难以达到工程要求精度的现状,利用过程神经网络输入为时间函数以及预测精度高的特点,建立了基于过程神经网络的电力系统短期负荷预测模型;给出了模型的结构,基于函数正交基展开的离散数据拟合方法以及模型的学习算法.针对东北某地区电网的日负荷数据,进行了模型训练和负荷预测正确性的研究.结果表明,所建立的预测模型对负荷的预测准确率高,优于BP神经网络负荷预测模型的预测结果.     

求解大型特征值问题的块Davidson方法的精化技术

块Davidson方法是求解大型对称矩阵特征值问题的一种有效的方法.但对一些特征值问题,当Ritz值收敛以后,该方法并不能保证Ritz向量也同时收敛.因此,为加速块Davidson方法的收敛性,研究了块Davidson方法的重新开始技术,提出了精化块Davidson方法,并对精化块Davidson方法进行了收敛性分析.数值试验和理论分析均表明,新方法对计算大型对称矩阵的一些极端特征对是有效的.     

射频磁控溅射透明ZnO薄膜的退火性质分析

采用射频(RF)磁控溅射法在玻璃衬底上制备了c轴择优取向的ZnO薄膜。对所制备的ZnO薄膜在空气气氛中进行不同温度(350~600℃)的退火处理。利用XRD研究退火对ZnO薄膜晶体性能和应力状态的影响;用扫描电子显微镜(SEM)观察薄膜的表面形貌;用分光光谱仪测试薄膜的透光率。研究表明,随退火温度的升高,ZnO薄膜(002)衍射峰强度不断增强,半高宽逐渐减小;ZnO薄膜中沿c轴方向存在着的张应力在500℃退火时得到松弛;退火处理后薄膜的平均透光率变化不大,但透射光谱出现了“红移”现象。     

No communication without manipulation: A causal-deflationary view of information

In this paper, we shall describe a new account of information in communicational contexts, namely, a causal-deflationary one. Our approach draws from Timpson's deflationary view and supplies the field of philosophy of information with new tools that will help to clarify the underlying structure of communication: information is an abstract entity that must be involved in a causal link in order to achieve communication. In light of our account, communication is not merely the existence of statistical correlations between source and receiver, as usually understood from a purely formal view. Instead, communication is an asymmetric phenomenon involving causal notions: the destination system must be able to be causally manipulated by intervening on the source for successful communication. In a nutshell, we shall support the following lemma: no communication without manipulation.     

探讨积分的极限

证明了若可积函数列{fn}在[a,b]上一致收敛,则nl→im∞∫abfn(x)dx中极限运算与积分运算可交换,从而揭示了"积分的极限"解法的内在本质,并且对于limn→∞∫01xnF(x)dx及nl→im∞∫ab[f(x)]ndx两种类型给出了更为具体有效的一般性解法.     

基于加权主成分分析和改进密度峰值聚类的协同训练算法

[目的]针对协同训练算法在视图分割时未考虑噪声影响和两视图分类器对无标记样本标注不一致问题,提出了基于加权主成分分析和改进密度峰值聚类的协同训练算法.[方法]首先引入加权主成分分析对数据进行预处理,通过寻求初始有标记样本中特征和类标记之间的依赖关系求得各特征加权系数,再对加权变换后的数据进行降维并提取高贡献度特征进行视...     

全数字突发模式8PSK系统的时钟和载波同步

突发模式传送系统中前导字作为系统开销降低了数据传送效率,提出一种全数字无前导字突发模式8PSK接收机方案,接收机的时钟和载波恢复均采用前向同步技术,并对载波频偏恢复算法进行了详细推导,并通过仿真验证了算法的可行性和有效性。     

Direct inhibition of phospholipid scrambling activity in erythrocytes by potassium ions

The exposure of phosphatidylserine (PS) at the cell surface plays a critical role in blood coagulation and serves as a macrophage recognition moiety for the engulfment of apoptotic cells. Previous observations have shown that a high extracellular [K⁺] and selective K⁺ channel blockers inhibit PS exposure in platelets and erythrocytes. Here we show that the rate of PS exposure in erythrocytes decreases by ~50% when the intracellular [K⁺] increases from 0 to physiological concentrations. Using resealed erythrocyte membranes, we further show that lipid scrambling is inducible by raising the intracellular [Ca²⁺] and that K⁺ ions have a direct inhibitory effect on this process. Lipid scrambling in resealed ghosts occurs in the absence of cell shrinkage and microvesicle formation, processes that are generally attributed to Ca²⁺-induced lipid scrambling in intact erythrocytes. Thus, opening of Ca²⁺-sensitive K⁺ channels causes loss of intracellular K⁺ that results in reduced intrinsic inhibitory effect of these ions on scramblase activity. Received 11 September 2008; received after revision 17 October 2008; accepted 27 October 2008