Spaces of matrices of equal rank

Let M_m,n(F) denote the space of all mXn matrices over the algebraically closed field F. A subspace of M_m,n(F), all of whose nonzero elements have rank k, is said to be essentially decomposable if there exist nonsingular mXn matrices U and V respectively such that for any element A, UAV has the form

$\text{UAV=}\text{A}{}_1\text{A}{}_2\text{A}{}_3\text{0}$

where A₁ is iX(k–i) for some i?k. Theorem: If $\text{K}$ is a space of rank k matrices, then either $\text{K}$ is essentially decomposable or dim $\text{K}$?k+1. An example shows that the above bound on non-essentially-decomposable spaces of rank k matrices is sharp whenever n?2k–1.     

Linear maps preserving M–P inverses of matrices between matrix spaces over fields

Suppose F is a field of characteristic not 2. Let n and m be two arbitrary positive integers with n≥2. We denote by M _n (F) and S _n (F) the space of n×n full matrices and the space of n×n symmetric matrices over F, respectively. All linear maps from S _n (F) to M _m (F) preserving M–P inverses of matrices are characterized first, and thereby all linear maps from S _n (F) (M _n (F)) to S _m (F) (M _m (F)) preserving M–P inverses of matrices are characterized, respectively.     

Indecomposable cones

We study some relations between a reproducing cone K in a linear space V over a fully ordered field F and the cone Γ(K) in Hom (V,V) consisting of all operators A such that AK ? K. In particular, indecomposable cones are considered.     

Linear maps that preserve singular and nonsingular matrices

Let M (n,K) be the algebra of n × n matrices over an algebraically closed field K and T:M (n,K)→M (n,K) a linear transformation with the property that T maps nonsingular (singular) matrices to nonsingular (singular) matrices. Using some elementary facts from commutative algebra we show that T is nonsingular and maps singular matrices to singular matrices (T is nonsingular or T maps all matrices to singular matrices). Using these results we obtain Marcus and Moyl's characterization [T(x) = UXVorU^tXV for fixed U and V] from a result of Dieudonné's. Examples are given to show the hypothesis of algebraic closure in necessary.     

Maps preserving square-zero matrices on the algebra consisting of all upper triangular matrices

Let F be a field, T _n (F) (respectively, N _n (F)) the matrix algebra consisting of all n × n upper triangular matrices (respectively, strictly upper triangular matrices) over F. A ∈ T _n (F) is said to be square zero if A ² = 0. In this article, we firstly characterize non-singular linear maps on N _n (F) preserving square-zero matrices in both directions, then by using it we determine non-singular linear maps on T _n (F) preserving square-zero matrices in both directions.     

On the matrices of given rank in a large subspace

Let V be a linear subspace of M_n,p(K) with codimension lesser than n, where K is an arbitrary field and n?p. In a recent work of the author, it was proven that V is always spanned by its rank p matrices unless n=p=2 and K?F₂. Here, we give a sufficient condition on codim V for V to be spanned by its rank r matrices for a given r∈?1,p-1?. This involves a generalization of the Gerstenhaber theorem on linear subspaces of nilpotent matrices.     

On cone orderings and the linear complementarity problem

This paper first generalizes a characterization of polyhedral sets having least elements, which is obtained by Cottle and Veinott [6], to the situation in which Euclidean space is partially ordered by some general cone ordering (rather than the usual ordering). We then use this generalization to establish the following characterization of the class $\text{C}$ of matrices ($\text{C}$ arises as a generalization of the class of Z-matrices; see [4], [13], [14]): M∈$\text{C}$ if and only if for every vector q for which the linear complementarity problem (q,M) is feasible, the problem (q,M) has a solution which is the least element of the feasible set of (q,M) with respect to a cone ordering induced by some simplicial cone. This latter result generalizes the characterizations of K-and Z-matrices obtained by Cottle and Veinott [6] and Tamir [21], respectively.     

Positive operators on the n-dimensional ice cream cone

Let $\text{K}{}_{\mathrm{n}}\text{= {x ? R}{}^{\mathrm{n}}\text{: (x}{}_1{}^2\text{+ · +x}{}^2{}_{\mathrm{n?1}}\text{)}\text{1}\text{2}\text{? x}{}_{\mathrm{n}}\text{}}$ be the n-dimensional ice cream cone, and let Γ(K_n) be the cone of all matrices in $\text{R}$ⁿⁿ mapping K_n into itself. We determine the structure of Γ(K_n), and in particular characterize the extreme matrices in Γ(K_n).     

The family of D-weights for a two-rooted graph

Let Π(G) be the set of paths of a particular class Π from the initial to the terminal root of a two-rooted (possibly directed) graph G. We consider the family of $\text{D}$-weights defined by

where Π_x(G) is the family of subsets of Π(G) which cover x(G), the vertex set or the edge (arc) set of G.A number of the common properties and interrelations of these weights are discussed. Some of the weights have been considered previously, [1, 2], in the context of percolation theory but here only combinatorial arguments are used.     

Isotone functions,dual cones,and networks

If the collection of all real-valued functions defined on a finite partially ordered set S of n elements is identified in the natural way with Rⁿ, it is obvious that the subset of functions that are isotone or order preserving with respect to the given partial order constitutes a closed, convex, polyhedral cone K in Rⁿ. The dual cone K^* of K is the set of all linear functionals that are nonpositive of K. This article identifies the important geometric properties of K, and characterizes a nonredundant set of defining equations and inequalities for K^* in terms of a special class of partitions of S into upper and lower sets. These defining constraints immediately imply a set of extreme rays spanning K and K^*. One of the characterizations of K^* involves feasibility conditions on flows in a network. These conditions are also used as a tool in analysis.     

Mappings of subspaces into subsets

Let V_n(q) denote the n-dimensional vector space over the finite field with q elements, and L_n(q) be the lattice of subspaces of V_n(q). Two rank- and order-preserving maps from L_n(q) onto the lattice of subsets of an n-set are constructed. Three equivalent formulations of these maps are given: an inductive procedure based on an elementary combinatorial interpretation of a well-known pair of difference equations satisfied by the Gaussian coefficients [$\text{n}\text{k}$], a direct set-theoretical definition, and, a direct definition involving a certain pair of modular chains in L_n(q). The direct set-theoretical definition of one of these maps has already been given by Knuth. Knuth's map, however, may be systematically discovered by means of the inductive procedure and the direct lattice-theoretic definition shows how it can be generalized. As a further application of the pair of difference equations satisfied by [$\text{n}\text{k}$], a direct-combinatorial proof of an identity of Carlitz that expands Gaussian coefficients in terms of binomial coefficients has been formulated.     

On the duality operator of a convex cone

Let C be a convex set in Rⁿ. For each y?C, the cone of C at y, denoted by cone(y, C), is the cone {α(x ? y): α ? 0 and x?C}. If K is a cone in Rⁿ, we shall denote by $\text{K}{}^1$ its dual cone and by F(K) the lattice of faces of K. Then the duality operator of K is the mapping $\text{d}{}_{\mathrm{K}}\text{: F(K) → F(K}{}^1\text{)}$ given by $\text{d}{}_{\mathrm{K}}\text{(F) = (}\text{span}\text{F)}{}^{\mathrm{\perp }}\text{∩ K}{}^1$. Properties of the duality operator d_K of a closed, pointed, full cone K have been studied before. In this paper, we study d_K for a general cone K, especially in relation to d_{cone(y, K)}, where y?K. Our main result says that, for any closed cone K in Rⁿ, the duality operator d_K is injective (surjective) if and only if the duality operator d_{cone(y, K)} is injective (surjective) for each vector y?K ～ [K ∩ (? K)]. In the last part of the paper, we obtain some partial results on the problem of constructing a compact convex set C, which contains the zero vector, such that cone (0, C) is equal to a given cone.     

On linear preservers of normal maps

We study group induced cone (GIC) orderings generating normal maps. Examples of normal maps cover, among others, the eigenvalue map on the space of n × n Hermitian matrices as well as the singular value map on n × n complex matrices. In this paper, given two linear spaces equipped with GIC orderings induced by groups of orthogonal operators, we investigate linear operators preserving normal maps of the orderings. A characterization of the preservers is obtained in terms of the groups. The result is applied to show that the normal structure of the spaces is preserved under the action of the operators. In addition, examples are given.     

Linear maps that preserve matrices annihilated by a polynomial

Let M_n be the algebra of n×n matrices over an algebraically closed field of characteristic zero. Let f(x) be a polynomial over $\text{F}$ with at least two distinct roots. Then all nonsingular linear maps L:M_n→M_n that map matrix roots of F(x)=0 into matrix roots of f(x}=0 are found.     

A note on k-plane integral transforms

Let Π be a k-dimensional subspace of Rⁿ, n ? 2, and write x = (x′, x″) with x′ in Π and x″ in the orthogonal complement Π^⊥. The k-plane transform of a measurable function ? in the direction Π at the point x″ is defined by $\text{L?(Π, x″) = ∝Π?(x′, x″) dx′}$. In this article certain a priori inequalities are established which show in particular that if $\text{? ? L}{}^{\mathrm{p}}\text{(R}{}^{\mathrm{n}}\text{)}$, $\text{1 ? p}\text{\$}\text{?}\text{n}\text{k}$, then ? is integrable over almost every translate of almost every k-space. Mapping properties of the k-plane transform between the spaces L^p(Rⁿ), p ? 2, and certain Lebesgue spaces with mixed norm on a vector bundle over the Grassmann manifold of k-spaces in Rⁿ are also obtained.     

On the matrix ring over a finite field

Let F_n be the ring of n × n matrices over the finite field F; let o(F_n) be the number of elements in F_n, and s(F_n) be the number of singular matrices in F_n. We prove that $\text{o(F}{}_{\mathrm{n}}\text{)<s(F}{}_{\mathrm{n}}\text{)}{}^{\mathrm{<mtext>1+1</mtext><mtext>n(n-1)</mtext>}}$ if n ? 2, and if n = 2 and o(F) ? 3, then .     

To what extent is a large space of matrices not closed under product?

Let (K) be a field. Given an arbitrary linear subspace V of M_n(K) of codimension less than n-1, a classical result states that V generates the (K)-algebra M_n(K). Here, we strengthen this statement in three ways: we show that M_n(K) is spanned by the products of the form AB with (A,B)∈V²; we prove that every matrix in M_n(K) can be decomposed into a product of matrices of V; finally, when V is a linear perplane of M_n(K) and n>2, we show that every matrix in M_n(K) is a product of two elements of V.     

Twisted inner products and contraction inequalities on spaces of contravariant and covariant tensors

Given positive integers n and p, and a complex finite dimensional vector space V, we let S_n,p(V) denote the set of all functions from V×V×?×V-(n+p copies) to C that are linear and symmetric in the first n positions, and conjugate linear symmetric in the last p positions. Letting κ=min{n,p} we introduce twisted inner products, [·,·]_s,t,1?s,t?κ, on S_n,p(V), and prove the monotonicity condition [F,F]_s,t?[F,F]_u,v is satisfied when s?u?κ,t?v?κ, and F∈S_n,p(V). Using the monotonicity condition, and the Cauchy-Schwartz inequality, we obtain as corollaries many known inequalities involving norms of symmetric multilinear functions, which in turn imply known inequalities involving permanents of positive semidefinite Hermitian matrices. New tensor and permanental inequalities are also presented. Applications to partial differential equations are indicated.     

Remarks on Banach spaces of compact operators

Let E and F be Banach spaces. We generalize several known results concerning the nature of the compact operators K(E, F) as a subspace of the bounded linear operators L(E, F). The main results are: (1) If E is a c₀ or l_p (1 < p < ∞) direct sum of a family of finite dimensional Banach spaces, then each bounded linear functional on K(E) admits a unique norm preserving extension to L(E). (2) If F has the bounded approximation property there is an isomorphism of $\text{L(E, F)}\text{into}\text{K(E, F)}{}^7$ such that its restriction to K(E, F) is the canonical injection. (3) If E is infinite dimensional and if F contains a complemented copy of c₀, K(E, F) is not complemented in L(E, F).