CONVERGENCE OF CASCADE ALGORITHMS AND SMOOTHNESS OF REFINABLE DISTRIBUTIONS

In this paper, the author at first develops a method to study convergence of the cascade algorithm in a Banach space without stable assumption on the initial (see Theorem 2.1), and then applies the previous result on the convergence to characterizing compactly supported refinable distributions in fractional Sobolev spaces and Holder continuous spaces (see Theorems 3.1, 3.3, and 3.4). Finally the author applies the above characterization to choosing appropriate initial to guarantee the convergence of the cascade algorithm (see Theorem 4.2).     

带惩罚项与随机输入的BP神经网络在线梯度学习算法的收敛性

本文对三层BP神经网络中带有惩罚项的在线梯度学习算法的收敛性问题进行了研究,在网络训练每一轮开始执行之前,对训练样本随机进行重排,以使网络学习更容易跳出局部极小,文中给出了误差函数的单调性定理以及该算法的弱收敛和强收敛性定理。     

Finding finer functions for partially characterized proteins by protein-protein interaction networks

Based on high-throughput data, numerous algorithms have been designed to find functions of novel proteins. However, the effectiveness of such algorithms is currently limited by some fundamental factors, including (1) the low a-priori probability of novel proteins participating in a detailed function; (2) the huge false data present in high-throughput datasets; (3) the incomplete data coverage of functional classes; (4) the abundant but heterogeneous negative samples for training the algorithms; and (5) the lack of detailed functional knowledge for training algorithms. Here, for partially characterized proteins, we suggest an approach to finding their finer functions based on protein interaction sub-networks or gene expression patterns, defined in function-specific subspaces. The proposed approach can lessen the above-mentioned problems by properly defining the prediction range and functionally filtering the noisy data, and thus can efficiently find proteins’ novel functions. For thousands of yeast and human proteins partially characterized, it is able to reliably find their finer functions (e.g., the translational functions) with more than 90% precision. The predicted finer functions are highly valuable both for guiding the follow-up wet-lab validation and for providing the necessary data for training algorithms to learn other proteins.     

A parallel method for time discretization of parabolic equations based on Laplace transformation and quadrature

We consider the discretization in time of an inhomogeneous parabolicequation in a Banach space setting, using a representation ofthe solution as an integral along a smooth curve in the complexleft half-plane which, after transformation to a finite interval,is then evaluated to high accuracy by a quadrature rule. Thisreduces the problem to a finite set of elliptic equations withcomplex coefficients, which may be solved in parallel. The paperis a further development of earlier work by the authors, wherewe treated the homogeneous equation in a Hilbert space framework.Special attention is given here to the treatment of the forcingterm. The method is combined with finite-element discretizationin spatial variables.     

On computation of the greatest common divisor of several polynomials over a finite field

We propose a probabilistic algorithm to reduce computing the greatest common divisor of m polynomials over a finite field (which requires computing m−1 pairwise greatest common divisors) to computing the greatest common divisor of two polynomials over the same field.     

线性规划联合算法的理论与应用

本在[1]的基础上．较系统的叙述了线性规划联合算法的步骤、相关理论及其应用，指出该算法具有避免人工变量、减少迭代次数、使用灵活、应用方便等特点。     

M/G/1非空竭服务休假排队系统随机分解的简化算法

本文根据M／G／1非空竭服务休假排队系统稳态队长随机分解的结构特征提出一种统一算法，该方法简洁高效，避免了再生循环方法繁杂的运算。运用该方法得出的结果与已知的用再生循环方法得出的结论一致。并且修正了Levy(1989)关于Bernoulli闸门服务休假排队系统随机分解的一个错误。     

求解一类非单调线性互补问题的宽邻域内点方法及其计算复杂性

对于一类非单调线性互补问题给出了一种新的算法——宽邻域内点算法，并讨论了其计算复杂性。     

线性规划问题的规范型算法

提出了线性规划问题的两种规范标准形式；证明了任意一个线性规划问题都可化为这两种形式之一；给出了不需引入人工变量的线性规划问题的求解算法。     

资源有限的加权总完工时间单机排序问题

本讨论资源有限的加权总工时间单机排序问题，对现在仍为OPEN问题1｜pj=bj-ajuj，∑uj≤U｜∑wjCj给出了一个有关最优解中最优资源分配的重要性质，并利用该性质分别给出了三种情况bj=b,wj=w,aj=a；bj=b,wj=w,uj=u；aj=a,wj=w,uj=u的最优算法。