No label defined (Q4832228)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 2124906
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 2124906 |
Statements
4 January 2005
0 references
doubly stochastic operator
0 references
positive operator
0 references
Birkhoff's theorem
0 references
0.92222124
0 references
0.89621073
0 references
0.8941498
0 references
0 references
0.87173414
0 references
0.8705813
0 references
0.8704593
0 references
0.8683933
0 references
0.86703265
0 references
``Let \((t_{ij})\) be any matrix of non-negative terms from an ordered filed, such that sums of the terms in each row and column is one. Show that it is a weighted mean of permutation matrices.'' This is Exercise 4 on page 266 in the second edition of \textit{G. Birkhoff}'s monograph [``Lattice theory'' (AMS Colloq. Publ. 25, AMS, New York) (1948; Zbl 0033.10103)]. The problem 111 in that monograph is based on Exercise 4 and stated as follows. Problem 111: Extend the result of Exercise 3 to the infinite-dimensional case, under suitable hypotheses. The paper under review is a survey on the Problem 111, the so-called Birkhoff theorem. In the present paper, the study of Birkhoff's theorem is divided into three cases., namely, the finite case, the countable case and the uncountable case. Birkhoff's theorem for these three cases is stated as follows. For the finite case, an \(n\times n\) matrix \(P=(p_{ij})_{n\times n}\) is called a stochastic matrix if and only if \(\sum _{j=1}^np_{ij}=1\) and \(p_{ij}\geq 0\) for all \(1\leq i\leq n.\) Moreover, \(P\) is called a doubly stochastic matrix if and only if both \(P\) and \(P^t\) are stochastic. The set of all doubly stochastic matrices is denoted by \(D_n\) and the set of all permutation matrices is denoted by \(P_n\). Birkhoff's theorem for the finite case then says that \(D_n=\text{ch}\,P_n\), where \(\text{ch}\,K\) denotes the convex hull of the set \(K\). For the countably infinite case, Birkhoff's theorem says that there exists a topology on the space \(V\) such that \(D=\text{cch}\,P\), where \(V\) is the linear space of all boundedly line-summable infinite matrices, i.e., \(M=(m_{ij})\in V\) if and only if \[ \sup_i\sum_j\mid m_{ij}\mid <\infty \text{ and } \sup_j\sum_i\mid m_{ij}\mid<\infty, \] and \(D\subset V\) is the set of all doubly stochastic matrices, \(P\subset V\) is the set of all permutation matrices and \(\text{cch}\,P\) denotes the countably convex hull. For the uncountable case, Birkhoff's theorem says that \(D\) is the closed convex hull of \(P\), the permutation operators in the strong operator topology. Here, \(D\) is the set of all double stochastic operators and \(P\) is the set of all permutation operators. In the present survey, the authors give an accurate ``road map'' to study problem 111, beginning with the finite case and proceeding to the uncountable case. They point out that there are many differences among the three cases. In the last section, they present their conclusions and observations, together with comments and suggestions which should prove useful for further study. For the entire collection see [Zbl 1048.60003].
0 references