On a positive set theory with inequality

We introduce a quite natural Frege‐style set theory, which we call Strong‐Frege‐2 $(\mathsf {SF}_2)$, a sort of simplification of the theory considered in 13 (under the name Strong‐Frege‐3) and 1 (under the name F2). We give a model of a weaker variant of $\mathsf {SF}_2$, called $\mathsf {SF}_2\mathsf {AC}$, where atoms and coatoms are allowed. To construct the model we use an enumeration “almost without repetitions” of the Π¹₁ sets of natural numbers; such an enumeration can be obtained via a classical priority argument much in the style of 5 and 15 . © 2011 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim     

Inequalities of Hadamard Type for r-Convex Functions in Carnot Groups

For a Carnot group G,we establish the relationship between extended mean values and r-convexfunctions which is introduced in this paper,which is a class of inequalities of Hadamard type for r-convexfunction on G.     

Generalized Multiplicative Inequalities for Ideal Spaces

We study the problem of completely describing the domains that enjoy the generalized multiplicative inequalities of the embedding theorem type. We transfer the assertions for the Sobolev spaces L _p ¹() to the function classes that result from the replacement of L _p () with an ideal space of vector-functions. We prove equivalence of the functional and geometric inequalities between the norms of indicators and the capacities of closed subsets of . The most comprehensible results relate to the case of the rearrangement invariant ideal spaces.     

广义实正定矩阵的几个不等式

推广了文献[3]中的广义实正定矩阵的行列式不等式,同时给出了广义实正定矩阵的凸性不等式.     

完备流形上带位势的热核估计

在完备非紧流形上获得了关于带位势热方程正解的梯度估计；接着，利用测地线的技巧获得了Harnack不等式；进一步，建立了两个积分不等式，综合Harnack不等式获得了热核的上下界；最后，利用函数的结果来控制p形式的热核。     

Q和Q_B准则在Double设计中的应用

一阶回归模型通常用来从众多因子当中筛选出效应显著的因子,而Q和QB准则能够比较简单地从众多合格的拟合子模型中找出最优设计.因此本文主要探讨了极大模型为一阶回归模型时,初始设计d分别与Double设计d dd-d,完全反转设计d-d,设计dd11-dd22之间的关系,以及它们在Q和QB准则下的最优性关系.     

一个核含无理式的Hilbert型不等式

应用权系数方法,给出一个带有最佳常数因子且核含有无理式的Hilbert型不等式,同时给出其逆向不等式及其等价形式.     

偶阶中立型分布时滞微分不等式的正解

证明了具有分布时滞的偶阶中立型泛函微分不等式不存在最终正解的新的充分条件,所得结果推广和改进了相关文献中的条件.     

具有变时滞的离散双向联想记忆神经网络的全局指数稳定性

利用Lyapunov函数和线性矩阵不等式(LMI),给出了判定一类双向联想记忆(BAM)神经网络模型的指数稳定的充分性条件.该条件去掉了以往论文中所要求的激活函数单调,可微分的条件,而且所得结果利用里的工具易于检测.并举例说明本文结果的有效性.     

Benders decomposition for a class of variational inequalities

This paper examines Benders decomposition for a useful class of variational inequality (VI) problems that can model, e.g., economic equilibrium, games or traffic equilibrium. The dual of the given VI is defined. Benders decomposition of the original VI is derived by applying a Dantzig–Wolfe decomposition procedure to the dual of the given VI, and converting the dual forms of the Dantzig–Wolfe master and subproblems to their primal forms. The master problem VI includes a new cut at each iteration, with information from the latest subproblem VI, which is solved by fixing the “difficult” variables at values determined by the previous master problem. A scalar parameter called the convergence gap is calculated at each iteration; a negative value is equivalent to the algorithm making progress in that the last master problem solution is made infeasible by the new cut. Under mild conditions, the convergence gap approaches zero in the limit of many iterations. With a more restrictive condition that still admits many useful models, a zero value of the convergence gap implies that the master problem has found a solution of the VI. A small model of competitive equilibrium of three commodities in two regions serves as an illustration.