A lower bound on the performance of the quadratic discriminant function

The quadratic discriminant function is often used to separate two classes of points in a multidimensional space. When the two classes are normally distributed, this results in the optimum separation. In some cases however, the assumption of normality is a poor one and the classification error is increased. The current paper derives an upper bound for the classification error due to a quadratic decision surface. The bound is strict when the class means and covariances and the quadratic discriminant surface satisfy certain specified symmetry conditions.     

Randomized on-line scheduling on three processors

We consider a randomized on-line scheduling problem where each job has to be scheduled on any of m identical processors. The objective is to minimize the expected makespan. We show that the competitive ratio of any randomized algorithm for m=3 processors must be strictly greater than .     

Construction of the extremal function for a functional on the classH ω (n)

For a functional on the classH _ω ⁽ⁿ⁾ ,n≥3, we construct the extremal function on which the upper bound obtained by A. I. Stepanets is attained. Translated fromMatematicheskie Zametki, Vol. 61, No. 4, pp. 519–529, April, 1997. Translated by N. K. Kulman     

Contractivity results in ordered spaces. Applications to relative operator bounds and projections with norm one

This paper provides various “contractivity” results for linear operators of the form where C are positive contractions on real ordered Banach spaces X . If A generates a positive contraction semigroup in Lebesgue spaces , we show (M. Pierre's result) that is a “contraction on the positive cone ”, i.e. for all provided that .  We show also that this result is not true for 1 ⩽ . We give an extension of M. Pierre's result to general ordered Banach spaces X under a suitable uniform monotony assumption on the duality map on the positive cone . We deduce from this result that, in such spaces, is a contraction on for any positive projection C with norm 1. We give also a direct proof (by E. Ricard) of this last result if additionally the norm is smooth on the positive cone. For any positive contraction C on base‐norm spaces X (e.g. in real spaces or in preduals of hermitian part of von Neumann algebras), we show that for all where N is the canonical half‐norm in X . For any positive contraction C on order‐unit spaces X (e.g. on the hermitian part of a algebra), we show that is a contraction on . Applications to relative operator bounds, ergodic projections and conditional expectations are given.     

一种求解二次约束二次规划问题的自适应全局优化算法

为了更好地解决二次约束二次规划问题(QCQP), 本文基于分支定界算法框架提出了自适应线性松弛技术, 在理论上证明了这种新的定界技术对于解决(QCQP)是可观的。文中分支操作采用条件二分法便于对矩形进行有效剖分; 通过缩减技术删除不包含全局最优解的部分区域, 以加快算法的收敛速度。最后, 通过数值结果表明提出的算法是有效可行的。     

两台自私型机器上自私工件排序的PoA紧界

本文研究了两台自私型机器上有自私型工件的关于二元均衡的排序问题。对任意工件序列$L$, 证明了二元均衡排序的PoA的紧界为$\frac{8}{7}$。如果工件尺寸在区间$[1, r](r\ge1)$内, 得到了二元均衡排序的PoA的紧界为关于$r$的分段线性函数。     

New perturbation bounds for the subunitary polar factors

In this article, we present some new perturbation bounds for the (subunitary) unitary polar factors of the (generalized) polar decompositions. Two numerical examples are given to show the rationality and superiority of our results, respectively. In terms of the one-to-one correspondence between the weighted case and the non-weighted case, all these bounds can be applied to the weighted polar decomposition.