Computable representations for convex hulls of low-dimensional quadratic forms

Let ${\mathcal{C}}$ be the convex hull of points ${{\{{1 \choose x}{1 \choose x}^T \,|\, x\in \mathcal{F}\subset \Re^n\}}}$ . Representing or approximating ${\mathcal{C}}$ is a fundamental problem for global optimization algorithms based on convex relaxations of products of variables. We show that if n ≤ 4 and ${\mathcal{F}}$ is a simplex, then ${\mathcal{C}}$ has a computable representation in terms of matrices X that are doubly nonnegative (positive semidefinite and componentwise nonnegative). We also prove that if n = 2 and ${\mathcal{F}}$ is a box, then ${\mathcal{C}}$ has a representation that combines semidefiniteness with constraints on product terms obtained from the reformulation-linearization technique (RLT). The simplex result generalizes known representations for the convex hull of ${{\{(x_1, x_2, x_1x_2)\,|\, x\in\mathcal{F}\}}}$ when ${\mathcal{F}\subset\Re^2}$ is a triangle, while the result for box constraints generalizes the well-known fact that in this case the RLT constraints generate the convex hull of ${{\{(x_1, x_2, x_1x_2)\,|\, x\in\mathcal{F}\}}}$ . When n = 3 and ${\mathcal{F}}$ is a box, we show that a representation for ${\mathcal{C}}$ can be obtained by utilizing the simplex result for n = 4 in conjunction with a triangulation of the 3-cube.     

Helly-type theorems for intersections of sets starshaped via orthogonally convex paths

Let ${\mathcal{K}}$ be a family of simply connected sets in the plane. If every countable subfamily of ${\mathcal{K}}$ has an intersection that is starshaped via orthogonally convex paths, then ${\mathcal{K}}$ itself has such an intersection. For the d-dimensional case, let ${\mathcal{K}}$ be a family of compact sets in ${\mathbb{R}^d}$ . If every finite subfamily of ${\mathcal{K}}$ has an intersection that is starshaped via orthogonally convex paths, again ${\mathcal{K}}$ itself has such an intersection.     

Continuity of solutions of a problem in the calculus of variations

We study the problem of minimizing ${\int_{\Omega} L(x,u(x),Du(x))\,{\rm d}x}$ over the functions ${u\in W^{1,p}(\Omega)}$ that assume given boundary values ${\phi}$ on ???. We assume that L(x, u, Du)?=?F(Du)?+?G(x, u) and that F is convex. We prove that if ${\phi}$ is continuous and ?? is convex, then any minimum u is continuous on the closure of ??. When ?? is not convex, the result holds true if F(Du)?=?f(|Du|). Moreover, if ${\phi}$ is Lipschitz continuous, then u is H?lder continuous.     

On convex relaxations for quadratically constrained quadratic programming

We consider convex relaxations for the problem of minimizing a (possibly nonconvex) quadratic objective subject to linear and (possibly nonconvex) quadratic constraints. Let $\mathcal{F }$ denote the feasible region for the linear constraints. We first show that replacing the quadratic objective and constraint functions with their convex lower envelopes on $\mathcal{F }$ is dominated by an alternative methodology based on convexifying the range of the quadratic form $\genfrac(){0.0pt}{}{1}{x}\genfrac(){0.0pt}{}{1}{x}^T$ for $x\in \mathcal{F }$ . We next show that the use of ?? $\alpha $ BB?? underestimators as computable estimates of convex lower envelopes is dominated by a relaxation of the convex hull of the quadratic form that imposes semidefiniteness and linear constraints on diagonal terms. Finally, we show that the use of a large class of D.C. (??difference of convex??) underestimators is dominated by a relaxation that combines semidefiniteness with RLT constraints.     

On additive decompositions of the set of primitive roots modulo p

It is conjectured that the set ${\mathcal {G}}$ of the primitive roots modulo p has no decomposition (modulo p) of the form ${\mathcal {G}= \mathcal {A} +\mathcal {B}}$ with ${|\mathcal {A}|\ge 2}$ , ${|\mathcal {B} |\ge 2}$ . This conjecture seems to be beyond reach but it is shown that if such a decomposition of ${\mathcal {G}}$ exists at all, then ${|\mathcal {A} |}$ , ${|\mathcal {B} |}$ must be around p ^1/2, and then this result is applied to show that ${\mathcal {G}}$ has no decomposition of the form ${\mathcal {G} =\mathcal {A} + \mathcal {B} + \mathcal {C}}$ with ${|\mathcal {A} |\ge 2}$ , ${|\mathcal {B} |\ge 2}$ , ${|\mathcal {C} |\ge 2}$ .     

A structural characterization of numéraires of convex sets of nonnegative random variables

We introduce the concept of numéraire s of convex sets in ${L^0_{+}}$ , the nonnegative orthant of the topological vector space L ⁰ of all random variables built over a probability space. A necessary and sufficient condition for an element of a convex set ${\mathcal{C} \subseteq L^0_{+}}$ to be a numéraire of ${\mathcal{C}}$ is given, inspired from ideas in financial mathematics.     

Boundedness and Compactness of Pseudodifferential Operators with Non-Regular Symbols on Weighted Lebesgue Spaces

Applying the boundedness on weighted Lebesgue spaces of the maximal singular integral operator S _* related to the Carleson?CHunt theorem on almost everywhere convergence, we study the boundedness and compactness of pseudodifferential operators a(x, D) with non-regular symbols in ${L^\infty(\mathbb{R}, V(\mathbb{R})), PC(\overline{\mathbb{R}}, V(\mathbb{R}))}$ and ${\Lambda_\gamma(\mathbb{R}, V_d(\mathbb{R}))}$ on the weighted Lebesgue spaces ${L^p(\mathbb{R},w)}$ , with 1?< p <? ?? and ${w\in A_p(\mathbb{R})}$ . The Banach algebras ${L^\infty(\mathbb{R}, V(\mathbb{R}))}$ and ${PC(\overline{\mathbb{R}}, V(\mathbb{R}))}$ consist, respectively, of all bounded measurable or piecewise continuous ${V(\mathbb{R})}$ -valued functions on ${\mathbb{R}}$ where ${V(\mathbb{R})}$ is the Banach algebra of all functions on ${\mathbb{R}}$ of bounded total variation, and the Banach algebra ${\Lambda_\gamma(\mathbb{R}, V_d(\mathbb{R}))}$ consists of all Lipschitz ${V_d(\mathbb{R})}$ -valued functions of exponent ${\gamma \in (0,1]}$ on ${\mathbb{R}}$ where ${V_d(\mathbb{R})}$ is the Banach algebra of all functions on ${\mathbb{R}}$ of bounded variation on dyadic shells. Finally, for the Banach algebra ${\mathfrak{A}_{p,w}}$ generated by all pseudodifferential operators a(x, D) with symbols ${a(x, \lambda) \in PC(\overline{\mathbb{R}}, V(\mathbb{R}))}$ on the space ${L^p(\mathbb{R}, w)}$ , we construct a non-commutative Fredholm symbol calculus and give a Fredholm criterion for the operators ${A \in \mathfrak{A}_{p,w}}$ .     

Stability of contact discontinuity for steady Euler system in infinite duct

In this paper, we prove stability of contact discontinuities for full Euler system. We fix a flat duct ${\mathcal{N}_0}$ of infinite length in ${\mathbb{R}^2}$ with width W ₀ and consider two uniform subsonic flow ${{U_l}^{\pm}=(u_l^{\pm}, 0, pl,\rho_l^{\pm})}$ with different horizontal velocity in ${\mathcal{N}_0}$ divided by a flat contact discontinuity ${\Gamma_{cd}}$ . And, we slightly perturb the boundary of ${\mathcal{N}_0}$ so that the width of the perturbed duct converges to ${W_0+\omega}$ for ${|\omega| < \delta}$ at ${x=\infty}$ for some ${\delta >0 }$ . Then, we prove that if the asymptotic state at left far field is given by ${{U_l}^{\pm}}$ , and if the perturbation of boundary of ${\mathcal{N}_0}$ and ${\delta}$ is sufficiently small, then there exists unique asymptotic state ${{U_r}^{\pm}}$ with a flat contact discontinuity ${\Gamma_{cd}^*}$ at right far field( ${x=\infty}$ ) and unique weak solution ${U}$ of the Euler system so that U consists of two subsonic flow with a contact discontinuity in between, and that U converges to ${{U_l}^{\pm}}$ and ${{U_r}^{\pm}}$ at ${x=-\infty}$ and ${x=\infty}$ respectively. For that purpose, we establish piecewise C ¹ estimate across a contact discontinuity of a weak solution to Euler system depending on the perturbation of ${\partial\mathcal{N}_0}$ and ${\delta}$ .     

On the Equality Problem of Conjugate Means

Let ${I\subset\mathbb{R}}$ be a nonvoid open interval and let L : I ²→ I be a fixed strict mean. A function M : I ²→ I is said to be an L-conjugate mean on I if there exist ${p,q\in\,]0,1]}$ and ${\varphi\in CM(I)}$ such that $$M(x,y):=\varphi^{-1}(p\varphi(x)+q\varphi(y)+(1-p-q) \varphi(L(x,y)))=:L_\varphi^{(p,q)}(x,y),$$ for all ${x,y\in I}$ . Here L(x, y) : = A _χ(x, y) ${(x,y\in I)}$ is a fixed quasi-arithmetic mean with the fixed generating function ${\chi\in CM(I)}$ . We examine the following question: which L-conjugate means are weighted quasi-arithmetic means with weight ${r\in\, ]0,1[}$ at the same time? This question is a functional equation problem: Characterize the functions ${\varphi,\psi\in CM(I)}$ and the parameters ${p,q\in\,]0,1]}$ , ${r\in\,]0,1[}$ for which the equation $$L_\varphi^{(p,q)}(x,y)=L_\psi^{(r,1-r)}(x,y)$$ holds for all ${x,y\in I}$ .     

A strict minimax inequality criterion and some of its consequences

In this paper, we point out a very flexible scheme within which a strict minimax inequality occurs. We then show the fruitfulness of this approach presenting a series of various consequences. Here is one of them: Let Y be a finite-dimensional real Hilbert space, J : Y ?? R a C ¹ function with locally Lipschitzian derivative, and ${\varphi : Y \to [0, + \infty[}$ a C ¹ convex function with locally Lipschitzian derivative at 0 and ${\varphi^{-1}(0) = \{0\}}$ . Then, for each ${x_0 \in Y}$ for which J??(x ₀)??? 0, there exists ???> 0 such that, for each ${r \in ]0, \delta[}$ , the restriction of J to B(x ₀, r) has a unique global minimum u _r which satisfies $$J(u_r)\leq J(x)-\varphi(x-u_r)$$ for all ${x \in B(x_0, r)}$ , where ${B(x_0, r) = \{x \in Y : \|x-x_0\|\leq{r}\}.}$      

New approximations for the cone of copositive matrices and its dual

We provide convergent hierarchies for the convex cone $\mathcal{C }$ of copositive matrices and its dual $\mathcal{C }^*$ , the cone of completely positive matrices. In both cases the corresponding hierarchy consists of nested spectrahedra and provide outer (resp. inner) approximations for $\mathcal{C }$ (resp. for its dual $\mathcal{C }^*$ ), thus complementing previous inner (resp. outer) approximations for $\mathcal{C }$ (for $\mathcal{C }^*$ ). In particular, both inner and outer approximations have a very simple interpretation. Finally, extension to $\mathcal{K }$ -copositivity and $\mathcal{K }$ -complete positivity for a closed convex cone $\mathcal{K }$ , is straightforward.     

Mixed Invariant Subspaces over the Bidisk

It is known that the structure of invariant subspaces I of the Hardy space H ² over the bidisk is extremely complicated. One reason is that it is difficult to describe infinite dimensional wandering spaces ${I\ominus zI}$ completely. In this paper, we study the structure of nontrivial closed subspaces N of H ² with ${T_zN\subset N}$ and ${T^*_wN\subset N}$ , which are called mixed invariant subspaces under T _z and ${T^*_w}$ . We know that the dimension of ${N\ominus zN}$ ranges from 1 to ??. If ${T^*_w(N\ominus zN)\subset N\ominus zN}$ , we may describe N completely. If ${T^*_w(N\ominus zN)\not\subset N\ominus zN}$ , it seems difficult to describe N generally. So we study N under the condition ${dim\,(N\ominus zN)=1}$ . Write ${M=H^2\ominus N}$ . We describe ${M\ominus wM}$ precisely. We give a characterization of N for which there is a nonzero function ${\varphi}$ in ${M\ominus wM}$ satisfying ${z^k\varphi\in M\ominus wM}$ for every k ?? 0. We also see that the space ${M\ominus wM}$ has a deep connection with the de Branges?CRovnyak spaces studied by Sarason.     

On the L p -distortion of finite quotients of amenable groups

We study the L ^p -distortion of finite quotients of amenable groups. In particular, for every ${2\leq p < \infty}$ , we prove that the ? ^p -distortions of the groups ${C_2\wr C_n}$ and ${C_{2^n}\rtimes C_n}$ are in ${\Theta((\log n)^{1/p}),}$ and that the ? ^p -distortion of ${C_n^2 \rtimes_A \mathbf{Z}}$ , where A is the matrix ${{\left({\small\begin{array}{cc}2 & 1 \\ 1 & 1 \end{array}} \right)}}$ is in ${\Theta((\log \log n)^{1/p}).}$      

Separable forms of {{\mathbb{G}}_m} -ACTIONS ON {{\mathbb{A}}^3}

Let k be a field of characteristic zero. We consider k-forms of $ {\mathbb G} $ _m -actions on $ {\mathbb A} $ ³ and show that they are linearizable. In particular, $ {\mathbb G} $ _m -actions on $ {\mathbb A} $ ³ are linearizable, and k-forms of $ {\mathbb A} $ ³ that admit an effective action of an infinite reductive group are trivial.     

A Question from a Famous Paper of Erdős

Given a convex body $K$ K , consider the smallest number $N$ N so that there is a point $P\in \partial K$ P ∈ ? K such that every circle centred at $P$ P intersects $\partial K$ ? K in at most $N$ N points. In 1946 Erd?s conjectured that $N=2$ N = 2 for all $K$ K , but there are convex bodies for which this is not the case. As far as we know there is no known global upper bound. We show that no convex body has $N=\infty $ N = ∞ and that there are convex bodies for which $N = 6$ N = 6 .     

An isoperimetric deficit upper bound of the convex domain in a surface of constant curvature

In this paper, we give a reverse analog of the Bonnesen-style inequality of a convex domain in the surface $ \mathbb{X} $ _∈ of constant curvature ∈, that is, an isoperimetric deficit upper bound of the convex domain in $ \mathbb{X} $ _∈ . The result is an analogue of the known Bottema’s result of 1933 in the Euclidean plane $ \mathbb{E} $ ².     

C 0 and bi-Lipschitz {\mathcal{K}} -equivalence of mappings

In this paper we investigate the classification of mappings up to ${\mathcal{K}}$ -equivalence. We give several results of this type. We study semialgebraic deformations up to semialgebraic C ⁰ ${\mathcal{K}}$ -equivalence and bi-Lipschitz ${\mathcal{K}}$ -equivalence. We give an algebraic criterion for bi-Lipschitz ${\mathcal{K}}$ -triviality in terms of semi-integral closure (Theorem 3.5). We also give a new proof of a result of Nishimura: we show that two germs of smooth mappings ${f, g: \mathbb{R}^n \to \mathbb{R}^n}$ , finitely determined with respect to ${\mathcal{K}}$ -equivalence are C ⁰- ${\mathcal{K}}$ -equivalent if and only if they have the same degree in absolute value.     

Varieties Generated by Wreath Products of Abelian Groups

This paper is a survey of our recent results concerning metabelian varieties, and more specifically, varieties generated by wreath products of Abelian groups. We give a full classification of cases where sets of wreath products of Abelian groups $ \mathfrak{X} $ Wr $ \mathfrak{Y} $ = { X Wr Y | X ∈ $ \mathfrak{X} $ , Y ∈ $ \mathfrak{Y} $ } and $ \mathfrak{X} $ wr $ \mathfrak{Y} $ = {X wr Y | X ∈ $ \mathfrak{X} $ , Y ∈ $ \mathfrak{Y} $ } generate the product variety $ \mathfrak{X} $ var ( $ \mathfrak{Y} $ ).     

A note on the reducibility of binary affine polynomials

Stickelberger–Swan Theorem is an important tool for determining parity of the number of irreducible factors of a given polynomial. Based on this theorem, we prove in this note that every affine polynomial A(x) over ${\mathbb{F}_2}$ with degree >1, where A(x) = L(x) + 1 and ${L(x)=\sum_{i=0}^{n}{x^{2^i}}}$ is a linearized polynomial over ${\mathbb{F}_2}$ , is reducible except x ² + x + 1 and x ⁴ + x + 1. We also give some explicit factors of some special affine pentanomials over ${\mathbb{F}_2}$ .     

A unified algorithm for mixed l_{2,p}-minimizations and its application in feature selection

Recently, matrix norm $l_{2,1}$ has been widely applied to feature selection in many areas such as computer vision, pattern recognition, biological study and etc. As an extension of $l_1$ norm, $l_{2,1}$ matrix norm is often used to find jointly sparse solution. Actually, computational studies have showed that the solution of $l_p$ -minimization ( $0<p<1$ ) is sparser than that of $l_1$ -minimization. The generalized $l_{2,p}$ -minimization ( $p\in (0,1]$ ) is naturally expected to have better sparsity than $l_{2,1}$ -minimization. This paper presents a type of models based on $l_{2,p}\ (p\in (0, 1])$ matrix norm which is non-convex and non-Lipschitz continuous optimization problem when $p$ is fractional ( $0<p<1$ ). For all $p$ in $(0, 1]$ , a unified algorithm is proposed to solve the $l_{2,p}$ -minimization and the convergence is also uniformly demonstrated. In the practical implementation of algorithm, a gradient projection technique is utilized to reduce the computational cost. Typically different $l_{2,p}\ (p\in (0,1])$ are applied to select features in computational biology.