On a characterization of irreducibility of a non-negative matrix

Let A denote an n×n matrix with all its elements real and non-negative, and let r_i be the sum of the elements in the ith row of A, i=1,…,n. Let B=A?D(r₁,…,r_n), where D(r₁,…,r_n) is the diagonal matrix with r_i at the position (i,i). Then it is proved that A is irreducible if and only if rank B=n?1 and the null space of B^T contains a vector d whose entries are all non-null.     

Non-linear Approximations Using Sets of Finite Cardinality or Finite Pseudo-dimension

We investigate optimal non-linear approximations of multivariate periodic functions with mixed smoothness. In particular, we study optimal approximation using sets of finite cardinality (as measured by the classical entropy number), as well as sets of finite pseudo-dimension (as measured by the non-linear widths introduced by Ratsaby and Maiorov). Approximation error is measured in the L_q(T^d)-sense, where T^d is the d-dimensional torus. The functions to be approximated are in the unit ball SB^r_p, θ of the mixed smoothness Besov space or in the unit ball SW^r_p of the mixed smoothness Sobolev space. For 1<p, q<∞, 0<θ⩽∞ and r>0 satisfying some restrictions, we establish asymptotic orders of these quantities, as well as construct asymptotically optimal approximation algorithms. We particularly prove that for either r>1/p and θ⩾p or r>(1/p−1/q)₊ and θ⩾min{q, 2}, the asymptotic orders of these quantities for the Besov class SB^r_p, θ are both n^−r(log n)^{(d−1)(r+1/2−1/θ)}.     

On the representation of integers by p-adic diagonal forms

Let p be an odd prime, let d be a positive integer such that (d,p?1)=1, let r denote the p-adic valuation of d and let m=1+3+3²+…+3^r. It is shown that for every p-adic integer n the equation Σ_i=1^mX_i^d=n has a nontrivial p-adic solution. It is also shown that for all p-adic units a₁, a₂, a₃, a₄ and all p-adic integers n the equation Σ_i=1⁴a_iX_i^p=n has a nontrivial p-adic solution. A corollary to each of these results is that every p-adic integer is a sum of four pth powers of p-adic integers.     

Boundedness for pseudodifferential operators on multivariate α-modulation spaces

The α-modulation spaces M ^s,α _p,q (R ^d ), α∈[0,1], form a family of spaces that contain the Besov and modulation spaces as special cases. In this paper we prove that a pseudodifferential operator σ(x,D) with symbol in the Hörmander class S ^b _ρ,0 extends to a bounded operator σ(x,D):M ^s,α _p,q (R ^d )→M ^s-b,α _p,q (R ^d ) provided 0≤α≤ρ≤1, and 1<p,q<∞. The result extends the well-known result that pseudodifferential operators with symbol in the class S ^b _1,0 maps the Besov space B ^s _p,q (R ^d ) into B ^s-b _p,q (R ^d ).     

On the metric theory of nearest integer continued fractions

Letb _i ,i=1, 2, 3,..., be the partial denominators of the regular continued fraction expansion of a real numberx between 0 and 1. Then, as is well known, for almost allx, the proportion ofb _i 's fori=1, 2, 3,...,n, which take the valuer, tends to $$\frac{1}{{\ln 2}} \cdot \ln \left\{ {1 + \frac{1}{{r(r + 2)}}} \right\},$$ and the mean value off(b _i ) fori=1, 2, 3,...,n tends to $$\frac{1}{{\ln 2}}\sum\limits_1^\infty {f(r)\ln } \left\{ {1 + \frac{1}{{r(r + 2)}}} \right\},$$ asn tends to infinity. It is well known that the mean value of lnb _i equals lnK, whereK is Khintchine's constant. We compute the mean values off(b _i ) for some other functions, viz. 1/b _i and 1/b _i ² , and also estimate them for another type of continued fractions, Thenearest integer continued fraction forx. From these limits, the rate of growth of the denominators of the convergents of regular and of nearest integer continued fractions is deduced. Finally, an application is made to the two types of continued fraction expansions of √N, whereN is a non-square positive integer.     

Newton Polygons of L Functions Associated with Exponential Sums of Polynomials of Degree Four over Finite Fields

Let F_q be the finite field of q elements with characteristic p and F_q^m its extension of degree m. Fix a nontrivial additive character Ψ of F_p. If f(x₁,…, x_n)∈F_q[x₁,…, x_n] is a polynomial, then one forms the exponential sum S_m(f)=∑_{(x1,…,xn)∈(Fq^m)ⁿ}Ψ(Tr_Fq^m/Fp(f(x₁,…,x_n))). The corresponding L functions are defined by L(f, t)=exp(∑^∞_m=0S_m(f)t^m/m). In this paper, we apply Dwork's method to determine the Newton polygon for the L function L(f(x), t) associated with one variable polynomial f(x) when deg f(x)=4. As an application, we also give an affirmative answer to Wan's conjecture for the case deg f(x)=4.     

THE EXTREME POINTS OF SEVERAL CLASSES OF ANALYTIC FUNCTIONS

1Intr0ducti0nLetAden0tethesetofallfunctionsanalyticinA={z:Izl<1}.LetB={W:WEAandIW(z)l51}.Aisalocallyconvexlineaztop0l0gicalspacewithrespecttothetopologyofuniformconvergenceon`c0mpact8ubsetsofA-LetTh(c1,'tc.-1)={p(z):p(z)EA,Rop(z)>0,p(z)=1 clz czzz ' c.-lz"-l 4z" ',wherecl,',cn-1areforedcomplexconstants}.LetTh,.(b,,-..,b,-,)={p(z):P(z)'EAwithReP(z)>Oandp(z)=1 blz ' b.-lz"-l 4z" '-,wherebl,-'-jbu-1areffeedrealconstantsanddkarerealnumbersf0rk=n,n 1,'--}-LetTu(l1,'i'tI.-1)={…     

Nested matrices and the existence of least majorized elements

We characterize the matrices A for which X(b)={x∣x∈Rⁿ, x?0, Ax?b, σⁿ_i=1x_i=1} contains a least majorized element for all vectors b satisfying X(b)≠?.     

Counting the number of twin Niven numbers

Given an integer q≥2, we say that a positive integer is a q-Niven number if it is divisible by the sum of its digits in base q. Given an arbitrary integer r∈[2,2q], we say that (n,n+1,…,n+r−1) is a q-Niven r -tuple if each number n+i, for i=0,1,…,r−1, is a q-Niven number. We show that there exists a positive constant c=c(q,r) such that the number of q-Niven r-tuples whose leading component is <x is asymptotic to cx/(log x) ^r as x→∞. Research of J.M. De Koninck supported in part by a grant from NSERC. Research of I. Kátai supported by the Applied Number Theory Research Group of the Hungarian Academy of Science and by a grant from OTKA.     

On exact coverings of the integers

By an exact covering of modulusm, we mean a finite set of liner congruencesx≡a _i (modm _i ), (i=1,2,...r) with the properties: (I)m _i ∣m, (i=1,2,...,r); (II) Each integer satisfies precisely one of the congruences. Let α≥0, β≥0, be integers and letp andq be primes. Let μ (m) senote the Möbius function. Letm=p ^α q ^β and letT(m) be the number of exact coverings of modulusm. Then,T(m) is given recursively by $$\mathop \Sigma \limits_{d/m} \mu (d)\left( {T\left( {\frac{m}{d}} \right)} \right)^d = 1$$ .     

Intersection patterns of convex sets

LetK ₁,…K_n be convex sets inR ^d. For 0≦i denote byf _ithe number of subsetsS of {1,2,…,n} of cardinalityi+1 that satisfy ∩{K _i∶i∈S}≠Ø. We prove:Theorem.If f _d+r=0 for somer r>=0, then {fx161-1} This inequality was conjectured by Katchalski and Perles. Equality holds, e.g., ifK ₁=…=K_r=R^d andK _r+1,…,K_n aren?r hyperplanes in general position inR ^d. The proof uses multilinear techniques (exterior algebra). Applications to convexity and to extremal set theory are given.     

On the nonvanishing of homogeneous product sums

For integer n ≥ 1 let H_n = H_n(x, y, z) = Σ_{p + q + r = n}x^py^qz^r be the homogeneous product sum of weight n on three letters x, y, z. Morgan Ward conjectured that H_n ≠ 0 for all integers n, x, y, z with n > 1 and xyz ≠ 0. In support of this conjecture he proved that H_n ≠ 0 if n is even or if n + 2 is a prime number greater than 3. This paper adds considerably more evidence in support of Ward's conjecture by showing that in many cases H_n(a, b, c)¬=0 modulo 2, 4, or 16. The parity of H_n(a, b, c) is determined in all cases and, when H_n(a, b, c) is even, further congruences are given modulo 4 or 16.     

Perfect difference systems of sets and Jacobi sums

A perfect (v,{k_i∣1≤i≤s},ρ) difference system of sets (DSS) is a collection of s disjoint k_i-subsets D_i, 1≤i≤s, of any finite abelian group G of order v such that every non-identity element of G appears exactly ρ times in the multiset {a−b∣a∈D_i,b∈D_j,1≤i≠j≤s}. In this paper, we give a necessary and sufficient condition in terms of Jacobi sums for a collection {D_i∣1≤i≤s} defined in a finite field F_q of order q=ef+1 to be a perfect (q,{k_i∣1≤i≤s},ρ)-DSS, where each D_i is a union of cyclotomic cosets of index e (and the zero 0∈F_q). Also, we give numerical results for the cases e=2,3, and 4.     

Maximum IPP Codes of Length 3

Let Q be an alphabet with q elements. For any code C over Q of length n and for any two codewords a = (a ₁, . . . , a _n ) and b = (b ₁, . . . , b _n ) in C, let ${D({\bf a, b}) = \{(x_1, . . . , x_n) \in {Q^n} : {x_i} \in \{a_i, b_i\}\,{\rm for}\,1 \leq i \leq n\}}Let Q be an alphabet with q elements. For any code C over Q of length n and for any two codewords a = (a ₁, . . . , a _n ) and b = (b ₁, . . . , b _n ) in C, let D(a, b) = {(x₁, . . . , x_n) ? Qⁿ : x_i ? {a_i, b_i} for 1 ￡ i ￡ n}{D({\bf a, b}) = \{(x_1, . . . , x_n) \in {Q^n} : {x_i} \in \{a_i, b_i\}\,{\rm for}\,1 \leq i \leq n\}}. Let C^* = è_{a, b ? C}D(a, b){C^* = {{\bigcup}_{\rm {a,\,b}\in{C}}}D({\bf a, b})}. The code C is said to have the identifiable parent property (IPP) if, for any x ? C^*{{\rm {\bf x}} \in C^*}, ?_{x ? D(a, b)}{a, b} 1 ?{{\bigcap}_{{\rm x}{\in}D({\rm a,\,b})}\{{\bf a, b}\}\neq \emptyset} . Codes with the IPP were introduced by Hollmann et al [J. Combin. Theory Ser. A 82 (1998) 21–133]. Let F(n, q) = max{|C|: C is a q-ary code of length n with the IPP}.T? and Safavi-Naini [SIAM J. Discrete Math. 17 (2004) 548–570] showed that 3q + 6 - 6 é?{q+1}ù ￡ F(3, q) ￡ 3q + 6 - é6 ?{q+1}ù{3q + 6 - 6 \lceil\sqrt{q+1}\rceil \leq F(3, q) \leq 3q + 6 - \lceil 6 \sqrt{q+1}\rceil}, and determined F (3, q) precisely when q ≤ 48 or when q can be expressed as r ² + 2r or r ² + 3r +2 for r ≥ 2. In this paper, we establish a precise formula of F(3, q) for q ≥ 24. Moreover, we construct IPP codes of size F(3, q) for q ≥ 24 and show that, for any such code C and any x ? C^*{{\rm {\bf x}} \in C^*}, one can find, in constant time, a ? C{{\rm {\bf a}} \in C} such that if x ? D (c, d){{\rm {\bf x}} \in D ({\bf c, d})} then a ? {c, d}{{\rm {\bf a}} \in \{{\rm {\bf c, d}}\}}.     

Convex integer maximization via Graver bases

We present a new algebraic algorithmic scheme to solve convex integer maximization problems of the following form, where c is a convex function on R^d and w₁x,…,w_dx are linear forms on Rⁿ,

max{c(w₁x,…,w_dx):Ax=b,x∈Nⁿ}.     

On the Evaluation of Weil Sums of Dembowski-Ostrom Polynomials

Let F_q denote the finite field of q elements, q=p^e odd, let χ₁ denote the canonical additive character of F_q where χ₁(c)=e^2πiTr(c)/p for all c∈F_q, and let Tr represent the trace function from F_q to F_p. We are interested in evaluating Weil sums of the form S=S(a₁, …, a_n)=∑_x∈Fq χ₁(D(x)) where D(x)=∑ⁿ_i=1 a_ix^pα_i+pβ_i, α_i?β_i for each i, is known as a Dembowski-Ostrom polynomial (or as a D-O polynomial). Coulter has determined the value of S when D(x)=ax^pα+1; in this note we show how Coulter's methods can be generalized to determine the absolute value of S for any D-O polynomial. When e is even, we give a subclass of D-O polynomials whose Weil sums are real-valued, and in certain cases we are able to resolve the sign of S. We conclude by showing how Coulter's work for the monomial case can be used to determine a lower bound on the number of F^l_q-solutions to the diagonal-type equation ∑^l_i=1 x^pγ+1_i+(x_i+λ)^pγ+1=0, where l is even, e/gcd(γ, e) is odd, and h (X)=λ^pe−γX^pe−γ+λ^pγX is a permutation polynomial over F_q.     

Approximation of functions of finite variation by superpositions of a Sigmoidal function

The aim of this note is to generalize a result of Barron [1] concerning the approximation of functions, which can be expressed in terms of the Fourier transform, by superpositions of a fixed sigmoidal function. In particular, we consider functions of the type h(x) = ∫ℝ^d ƒ (〈t, x〉)dμ(t), where μ is a finite Radon measure on ℝ^d and ƒ : ℝ → ℂ is a continuous function with bounded variation in ℝ We show (Theorem 2.6) that these functions can be approximated in L₂-norm by elements of the set G_n = {Σ_i=0^staggeredn c_ig(〈a_i, x〉 + b_i) : a_iℝ^d, b_i, c_iℝ}, where g is a fixed sigmoidal function, with the error estimated by C/n^1/2, where C is a positive constant depending only on f. The same result holds true (Theorem 2.9) for f : ℝ → ℂ satisfying the Lipschitz condition under an additional assumption that ∫ℝ^d6t6e^d|u(t)| > ∞     

Nonnormality of Stoneham constants

This paper examines “Stoneham constants,” namely real numbers of the form $\alpha_{b,c} = \sum_{n \geq1} 1/(c^{n} b^{c^{n}})$ , for coprime integers b≥2 and c≥2. These are of interest because, according to previous studies, α _b,c is known to be b-normal, meaning that every m-long string of base-b digits appears in the base-b expansion of the constant with precisely the limiting frequency b ^?m . So, for example, the constant $\alpha_{2,3} = \sum_{n \geq1} 1/(3^{n} 2^{3^{n}})$ is 2-normal. More recently it was established that α _b,c is not bc-normal, so, for example, α _2,3 is provably not 6-normal. In this paper, we extend these findings by showing that α _b,c is not B-normal, where B=b ^p c ^q r, for integers b and c as above, p,q,r≥1, neither b nor c divide r, and the condition D=c ^q/p r ^1/p /b ^c?1<1 is satisfied. It is not known whether or not this is a complete catalog of bases to which α _b,c is nonnormal. We also show that the sum of two B-nonnormal Stoneham constants as defined above, subject to some restrictions, is B-nonnormal.     

The Monotonicity of Period Function for Codimension Four Quadratic System Q4

In this paper, we study the codimension four quadratic system Q₄ : ?=−iz+4z²+2∣z∣²+(b+ic)z?², where b and c are real constants, i²=−1, z=x+iy, ∣b+ic∣=2. It is proved that the period function of periodic trajectories for Q₄ is monotone. The content of this paper can be viewed as a contribution to the proof of Chicone's conjecture (cf. MR: 94h:58072).     

Spectrum of Ornstein-Uhlenbeck Operators in L Spaces with Respect to Invariant Measures

Let A=∑_i,j=1^Nq_ijD_ij+∑_i,j=1^Nb_ijx_jD_i be a possibly degenerate Ornstein-Uhlenbeck operator in R^N and assume that the associated Markov semigroup has an invariant measure μ. We compute the spectrum of A in L_μ^p for 1?p<∞.