Noncommutative convexity arises from linear matrix inequalities

This paper concerns polynomials in g noncommutative variables x=(x₁,…,xg), inverses of such polynomials, and more generally noncommutative “rational expressions” with real coefficients which are formally symmetric and “analytic near 0.” The focus is on rational expressions r=r(x) which are “matrix convex” near 0; i.e., those rational expressions r for which there is an ?>0 such that if X=(X₁,…,Xg) is a g-tuple of n×n symmetric matrices satisfying

     

Locally finite polynomial endomorphisms

We study polynomial endomorphisms F of C^N which are locally finite in the following sense: the vector space generated by r°Fⁿ (n≥0) is finite dimensional for each r∈C[x₁,…,x_N]. We show that such endomorphisms exhibit similar features to linear endomorphisms: they satisfy the Jacobian Conjecture, have vanishing polynomials, admit suitably defined minimal and characteristic polynomials, and the invertible ones admit a Dunford decomposition into “semisimple” and “unipotent” constituents. We also explain a relationship with linear recurrent sequences and derivations. Finally, we give particular attention to the special cases where F is nilpotent and where N=2.     

Infinitesimal non-crossing cumulants and free probability of type B

Free probabilistic considerations of type B first appeared in the paper of Biane, Goodman and Nica [P. Biane, F. Goodman, A. Nica, Non-crossing cumulants of type B, Trans. Amer. Math. Soc. 355 (2003) 2263-2303]. Recently, connections between type B and infinitesimal free probability were put into evidence by Belinschi and Shlyakhtenko [S.T. Belinschi, D. Shlyakhtenko, Free probability of type B: Analytic aspects and applications, preprint, 2009, available online at www.arxiv.org under reference arXiv:0903.2721]. The interplay between “type B” and “infinitesimal” is also the object of the present paper. We study infinitesimal freeness for a family of unital subalgebras A₁,…,Ak in an infinitesimal noncommutative probability space (A,φ,φ^′) and we introduce a concept of infinitesimal non-crossing cumulant functionals for (A,φ,φ^′), obtained by taking a formal derivative in the formula for usual non-crossing cumulants. We prove that the infinitesimal freeness of A₁,…,Ak is equivalent to a vanishing condition for mixed cumulants; this gives the infinitesimal counterpart for a theorem of Speicher from “usual” free probability. We show that the lattices NC(B)(n) of non-crossing partitions of type B appear in the combinatorial study of (A,φ,φ^′), in the formulas for infinitesimal cumulants and when describing alternating products of infinitesimally free random variables. As an application of alternating free products, we observe the infinitesimal analogue for the well-known fact that freeness is preserved under compression with a free projection. As another application, we observe the infinitesimal analogue for a well-known procedure used to construct free families of free Poisson elements. Finally, we discuss situations when the freeness of A₁,…,Ak in (A,φ) can be naturally upgraded to infinitesimal freeness in (A,φ,φ^′), for a suitable choice of a “companion functional” .     

When the positivity of the leading principal minors implies the positivity of all principal minors of a matrix

An n-by-n real matrix A enjoys the “leading implies all” (LIA) property, if, whenever D   is a diagonal matrix such that A+D$A+D$ has positive leading principal minors (PMs), all PMs of A are positive. Symmetric and Z-matrices are known to have this property. We give a new class of matrices (“mixed matrices”) that both unifies and generalizes these two classes and their special diagonal equivalences by also having the LIA property. “Nested implies all” (NIA) is also enjoyed by this new class.     

An Erd?s-Ko-Rado theorem for partial permutations

Let [n] denote the set of positive integers {1,2,…,n}. An r-partial permutation of [n] is a pair (A,f) where A⊆[n], |A|=r and f:A→[n] is an injective map. A set A of r-partial permutations is intersecting if for any (A,f), (B,g)∈A, there exists x∈A∩B such that f(x)=g(x). We prove that for any intersecting family A of r-partial permutations, we have .It seems rather hard to characterize the case of equality. For 8?r?n-3, we show that equality holds if and only if there exist x₀ and ε₀ such that A consists of all (A,f) for which x₀∈A and f(x₀)=ε₀.     

Global attractivity in a nonlinear second-order difference equation

In this paper we obtain a global attractivity result for the positive equilibrium of a nonlinear second-order difference equation of the form x_n+1 = f(x_n, x_n+1), n = 0, 1, ? The result applies to the difference equation x_n+1 =A+bx_n/A+_n?1, n = 0, 1, ? Where a, b, A ? (0, ∞). © 1996 John Wiley & Sons, Inc.     

On the span of Hadamard products of vectors

Let A₁, … , A_k be positive semidefinite matrices and B₁, … , B_k arbitrary complex matrices of order n. We show that

span{(A₁x)°(A₂x)°?°(A_kx)|x∈Cⁿ}=range(A₁°A₂°?°A_k)     

Green’s matrices

Our goal is to identify and understand matrices A that share essential properties of the unitary Hessenberg matrices M that are fundamental for Szegö’s orthogonal polynomials. Those properties include: (i) Recurrence relations connect characteristic polynomials {r_k(x)} of principal minors of A. (ii) A is determined by generators (parameters generalizing reflection coefficients of unitary Hessenberg theory). (iii) Polynomials {r_k(x)} correspond not only to A but also to a certain “CMV-like” five-diagonal matrix. (iv) The five-diagonal matrix factors into a product BC of block diagonal matrices with 2 × 2 blocks. (v) Submatrices above and below the main diagonal of A have rank 1. (vi) A is a multiplication operator in the appropriate basis of Laurent polynomials. (vii) Eigenvectors of A can be expressed in terms of those polynomials.Conditions (v) connects our analysis to the study of quasi-separable matrices. But the factorization requirement (iv) narrows it to the subclass of “Green’s matrices” that share Properties (i)-(vii).The key tool is “twist transformations” that provide 2ⁿ matrices all sharing characteristic polynomials of principal minors with A. One such twist transformation connects unitary Hessenberg to CMV. Another twist transformation explains findings of Fiedler who noticed that companion matrices give examples outside the unitary Hessenberg framework. We mention briefly the further example of a Daubechies wavelet matrix. Infinite matrices are included.     

Contributions to max-min convex geometry. I: Segments

We give some contributions to the theory of “max-min convex geometry”, that is, convex geometry in the semimodule over the max-min semiring R_max,min=R∪{-∞,+∞}. We introduce “elementary segments” that generalize from n=2 the horizontal, vertical or oblique segments contained in the main bisector of . We show that every segment in is a concatenation of a finite number of elementary subsegments (at most 2n-1, respectively at most 2n-2, in the case of comparable, respectively, incomparable, endpoints x,y). In this first part we study “max-min segments”, and in the subsequent second part (submitted) we study “max-min semispaces” and some of their relations to “max-min convex sets”.     

Group actions and Helly's theorem

We describe a connection between the combinatorics of generators for certain groups and the combinatorics of Helly's 1913 theorem on convex sets. We use this connection to prove fixed point theorems for actions of these groups on nonpositively curved metric spaces. These results are encoded in a property that we introduce called “property FAr”, which reduces to Serre's property FA when r=1. The method applies to S-arithmetic groups in higher Q-rank, to simplex reflection groups (including some nonarithmetic ones), and to higher rank Chevalley groups over polynomial and other rings (for example SLn(Z[x₁,…,xd]), n>2).     

Paley-Wiener spaces with vanishing conditions and Painlevé VI transcendents

We modify the classical Paley-Wiener spaces PWx of entire functions of finite exponential type at most x>0, which are square integrable on the real line, via the additional condition of vanishing at finitely many complex points z₁,…,zn. We compute the reproducing kernels and relate their variations with respect to x to a Krein differential system, whose coefficient (which we call the μ-function) and solutions have determinantal expressions. Arguments specific to the case where the “trivial zeros” z₁,…,zn are in arithmetic progression on the imaginary axis allow us to establish for expressions arising in the theory a system of two non-linear first order differential equations. A computation, having this non-linear system at his start, obtains quasi-algebraic and among them rational Painlevé transcendents of the sixth kind as certain quotients of such μ-functions.     

On the rational recursive sequence x_{n+1}=\frac{\alpha+\beta x_{n-k}}{\gamma-x_{n}}

In this paper, we investigate the global asymptotic stability, the periodicity nature and the boundedness character of the positive solutions of the difference equation x _n+1=(α+β x _n?k )/(γ?x _n ) where n=0,1,2,… and k∈{1,2,…}. The parameters α≥0, γ,β>0 and the initial conditions x _?k , x _?k+1,…,x _?1,x ₀ are real positive numbers. We show that the positive equilibrium point of this equation is a global attractor with a basin that depends on certain conditions posed on the coefficients α,β,γ.     

Interval partitions and Stanley depth

In this paper, we answer a question posed by Herzog, Vladoiu, and Zheng. Their motivation involves a 1982 conjecture of Richard Stanley concerning what is now called the Stanley depth of a module. The question of Herzog et al., concerns partitions of the non-empty subsets of {1,2,…,n} into intervals. Specifically, given a positive integer n, they asked whether there exists a partition P(n) of the non-empty subsets of {1,2,…,n} into intervals, so that |B|?n/2 for each interval [A,B] in P(n). We answer this question in the affirmative by first embedding it in a stronger result. We then provide two alternative proofs of this second result. The two proofs use entirely different methods and yield non-isomorphic partitions. As a consequence, we establish that the Stanley depth of the ideal (x₁,…,xn)⊆K[x₁,…,xn] (K a field) is ⌈n/2⌉.     

Asymptotic behavior of solutions of linear ordinary differential equations

We present some conditions which ensure that the solution Y(x) of the ordinary differential equation Y′(x) = A(x) Y(x), Y(x₀) = I, where x₀ ? x < ∞ and A(x), Y(x) are n × n complex matrix-valued functions with A(x) continuous, has a nonsingular limit as x → ∞.     

Linear combinations of Hermitian and real symmetric matrices

This paper, by purely algebraic and elementary methods, studies useful criteria under which the quadratic forms x′Ax and x′Bx, where A,B are n × n symmetric real matrices and x′=(x₁,x₂, …,x_n)≠(0,0,0,0, …,0), can vanish simultaneously and some real linear combination of A,B can be positive definite. Analogous results for hermitian matrices have also been discussed. We have given sufficient conditions on m real symmetric matrices so that some real linear combination of them can be positive definite.     

How to stay in a set or König’s lemma for random paths

Starting at statex?X, a player selects the next statex ₁ from the collection Τ(x) of those available and then selectsx ₂ from Τ(x ₁) and so on. Suppose the object is to control the pathx ₁,x ₂, … so that everyx _i will lie in a subsetA ofX. A famous lemma of König is equivalent to the statement that if every Τ(x) is finite and if, for everyn, the player can obtain a path inA of lengthn, then the player can obtain an infinite path inA. Here paths are not necessarily deterministic and, for eachx, Τ(x) is the collection of possible probability distributions for the next state. Under mild measurability conditions, it is shown that if, for everyn, there is a random path of lengthn which lies inA with probability larger than α, then there is an infinite random path with the same property. Furthermore, the measurability and finiteness assumptions can be dropped if, in the hypothesis, the positive integersn are replaced by stop rulest. An analogous result holds when the object is to visitA infinitely many times.     

Asymptotic estimates of sums involving the Moebius function

Let p(n) denote the smallest prime factor of an integer n>1 and let p(1)=∞. We study the asymptotic behavior of the sum M(x,y)=Σ_{1≤n≤x,p(n)>y}μ(n) and use this to estimate the size of A(x)=max_|f|≤1|Σ_2≤n<xμ(n)f(p(n))|, where μ(n) is the Moebius function. Applications of bounds for A(x), M(x,y) and similar quantities are discussed.     

Asymptotical Formulas For Solutions Of Linear Differential Systems Of The First Order

In this paper a system of differential equations y′ ? A(·,λ)y = 0 is considered on the finite interval [a,b] where λ ∈ C, A(·, λ):= λ A₁+ A ₀ +λ _?1A_?1(·,λ) and A ₁,A ₀, A _{? 1} are n × n matrix-functions. The main assumptions: A ₁ is absolutely continuous on the interval [a, b], A ₀ and A _{- 1}(·,λ) are summable on the same interval when ¦λ¦ is sufficiently large; the roots φ₁(x),…,φ_n (x) of the characteristic equation det (φ E — A ₁) = 0 are different for all x ∈ [a,b] and do not vanish; there exists some unlimited set Ω ? C on which the inequalities Re(λφ₁(x)) ≤ … ≤ Re (λφ_n(x)) are fulfilled for all x ∈ [a,b] and for some numeration of the functions φ_j(x). The asymptotic formula of the exponential type for a fundamental matrix of solutions of the system is obtained for sufficiently large ¦λ¦. The remainder term of this formula has a new type dependence on properties of the coefficients A ₁ (x), A _o (x) and A _{- 1} (x).     

EKR type inequalities for 4-wise intersecting families

Let 1?t?7 be an integer and let F be a k-uniform hypergraph on n vertices. Suppose that |A∩B∩C∩D|?t holds for all A,B,C,D∈F. Then we have if holds for some ε>0 and all n>n₀(ε). We apply this result to get EKR type inequalities for “intersecting and union families” and “intersecting Sperner families.”