On the asymptotic joint distribution of an unbounded number of sample extremes

Summary Convergence of the sample maximum to a nondegenerate random variable, as the sample sizen, implies the convergence in distribution of thek largest sample extremes to ak-dimensional random vectorM _k , for all fixedk. If we letk=k(n),k/n0, then a question arises in a natural way: how fast cank grow so that asymptotic probability statements are unaffected when sample extremes are replaced byM _k . We answer this question for two classes of events-the class of all Lebesgue sets inR ^k and the class of events of the form .     

A structure theorem for sum form functional equations

Summary It is proved that, iff _ij:]0, 1[ C (i = 1, ,k;j = 1, ,l) are measurable, satisfy the equation (1) (with some functionsg _it, h_jt:]0, 1[ C), then eachf _ij is in a linear space (called Euler space) spanned by the functionsx x ^j(logx) ^k (x ]0, 1[;j = 1, ,M;k = 0, ,m _j – 1), where ₁, , _M are distinct complex numbers andm ₁, , m_M natural numbers. The dimension of this linear space is bounded by a linear function ofN.     

Rational functions with partial quotients of small degree in their continued fraction expansion

For a rational functionf/g=f(x)/g(x) over a fieldF with ged (f,g)=1 and deg (g)1 letK(f/g) be the maximum degree of the partial quotients in the continued fraction expansion off/. ForfF[x] with deg (f)=k1 andf(O)O putL(f)=K(f(x)/x ^k ). It is shown by an explicit construction that for every integerb with 1bk there exists anf withL(f)=b. IfF=F ₂, the binary field, then for everyk there is exactly onefF ₂[x] with deg (f)=k,f(O)O, andL(f)=1. IfF _q is the finite field withq elements andgF _q [x] is monic of degreek1, then there exists a monic irreduciblefF _q [x] with deg (f)=k, gcd (f,g)=1, andK(f/g)<2+2 (logk)/logq, where the caseq=k=2 andg(x)=x ²+x+1 is excluded. An analogous existence theorem is also shown for primitive polynomials over finite fields. These results have applications to pseudorandom number generation.     

Matrix summability and a generalized Gibbs phenomenon

Letf be a real-valued function sequence {f _k } that converges to on a deleted neighborhoodD of . If there is a subsequence {f _k(j) } and a number sequencex such that lim _j x _j = and either lim _j f _k(j) (x _j )>lim sup _x (x) or lim _j f _k(j) (x _j ) x (x), thenf is said to display theGibbs phenomenon at . IfA is a (real) summability matrix, thenAf is a function sequence given by(Af) _n (x)= _k=0 a _n,k f _k (x). IfAf displays the Gibbs phenomenon wheneverf does, thenA is said to beGP-preserving. By replacingf _k (x) withf _k (x _j )F _k,j , the Gibbs phenomenon is viewed as a property of the matrixF, andGP-preserving matrices are determined by properties of the matrix productAF. The general results give explicit conditions on the entries {a _n,k } that are necessary and/or sufficient forA to beGP-preserving. For example: if(x)0 thenF displaysGP iff lim _k,j F _k,j 0, and ifA isGP-preserving then lim _n,k A _n,k 0. IfA is a triangular matrix that is stronger than convergence, thenA is notGP-preserving. The general results are used to study the preservation of the Gibbs phenomenon by matrix methods of Nörlund, Hausdorff, and others.     

On the exactness of certain inequalities in approximation theory

We prove the following: for every sequence {F_v}, F_v ? 0, F_v > 0 there exists a functionf such that

1. E_n(f)?F_n (n=0, 1, 2, ...) and
2. A_kn^?k? _v=1 ⁿ v^k?1 F_v?1?Ω_k (f, n^?1) (n=1, 2, ...).

     

Homotopies for computation of fixed points on unbounded regions

Givenf: R ⁿ R ^n* with some conditions, our aim is to compute a fixed pointx f(x) off; hereR ⁿ isn-dimensional Euclidean space andR ^n* is the collection of nonempty subsets ofR ⁿ . A typical application of the algorithm can be motivated as follows: Beginning with the constant mapf ₀:R ⁿ {0} R ⁿ and its fixed pointx ₀ = 0, we deformf _t ast tof f and follow the pathx _t of fixed points off _t . Cluster points of thex _t 's ast are fixed points off. This research was supported in part by Army Research Office-Durham Contract DAHC-04-71-C-0041 and by National Science Foundation Grant GK-5695.     

Convergence of the steepest descent method for minimizing quasiconvex functions

To minimize a continuously differentiable quasiconvex functionf: ⁿ , Armijo's steepest descent method generates a sequencex ^k+1 =x ^k –t _k f(x ^k ), wheret _k >0. We establish strong convergence properties of this classic method: either , s.t. ; or arg minf = , x ^k andf(x ^k ) inff. We also discuss extensions to other line searches.The research of the first author was supported by the Polish Academy of Sciences. The second author acknowledges the support of the Department of Industrial Engineering, Hong Kong University of Science and Technology.We wish to thank two anonymous referees for their valuable comments. In particular, one referee has suggested the use of quasiconvexity instead of convexity off.     

A state space isomorphism theorem for minimal dynamical systems

Consider a discrete time dynamical systemx _k+1=f(x _k ) on a compact metric spaceM, wheref:MM is a continuous map. Leth:MB ^k be a continuous output function. Suppose that all of the positive orbits off are dense and that the system is observable. We prove that any output trajectory of the system determinesf andh andM up to a homeomorphism. IfM is a compact Abelian topological group andf is an ergodic translation, then any output trajectory determines the system up to a translation and a group isomorphism of the group.     

On a Class of Operator Equations

We assume that E₁ and E₂ are Banach spaces, a:E₁ E₂ is a continuous linear surjective operator, f:E₁ E₂ is a nonlinear completely continuous operator. In this paper, we study existence problems for the equation a(x)=f(x) and estimate the topological dimension dim of the set of solutions.     

On the shape of the convex hull of random points

Summary Denote by E _n the convex hull of n points chosen uniformly and independently from the d-dimensional ball. Let Prob(d, n) denote the probability that E _n has exactly n vertices. It is proved here that Prob(d, 2 ^d/2 d ^-)1 and Prob(d, 2 ^d/2 d ^(3/4)+)0 for every fixed >0 when d. The question whether E _n is a k-neighbourly polytope is also investigated.     

Irrfahrten aufF 2

LetF ₂ be the free group with two generatorsa, b andP be a probability measure onF ₂ with support {a, a ^–1,b, b ^–1}. The asymptotic behavior forn of the convolution powersP ⁿ (e) at the identitye is found.

---

---

Herrn Prof. Dr. E. Hlawka zum 60. Geburtstag gewidmet     

Two combinatorial problems on posets

Bialostocki proposed the following problem: Let nk2 be integers such that k|n. Let p(n, k) denote the least positive integer having the property that for every poset P, |P|p(n, k) and every Z _k -coloring f: P Z _k there exists either a chain or an antichain A, |A|=n and _aA f(a) 0 (modk). Estimate p(n, k). We prove that there exists a constant c(k), depends only on k, such that (n+k–2)²–c(k) p(n, k) (n+k–2)²+1. Another problem considered here is a 2-dimensional form of the monotone sequence theorem of Erdös and Szekeres. We prove that there exists a least positive integer f(n) such that every integral square matrix A of order f(n) contains a square submatrix B of order n, with all rows monotone sequences in the same direction and all columns monotone sequences in the same direction (direction means increasing or decreasing).     

The reverse spelling of an FPrt-universal word in two letters

ForX a set the expression Prt(X) denotes the composition monoid of all functionsf X ×X. Fork a positive integer the letterk denotes also the set of all nonnegative integers less thank. Whenk > 1 the expression r_k denotes the connected injective element {<i, i + 1>i k – 1} in Prt (k). We show for every word w=w(x,y) in a two-letter alphabet that if the equation w(x, y)=r_k has a solution =y) ²Prt(k) then ¯w(x,y)=r_k also has a solution in²Prt(k), where ¯w is the word obtained by spelling the wordw backwards. It is a consequence of this theorem that if for every finite setX and for everyf Prt(X) the equation w(x,y)=f has a solution in²Prt(X) then for every suchX andf the equation ¯w(x, y)=f has a solution in²Prt(X).Presented by J. Mycielski.     

Asymptotically Optimal Tests and Optimal Designs for Testing the Mean in Regression Models with Applications to Change-Point Problems

Let a linear regression model be given with an experimental region [a, b] R and regression functions f ₁, ..., f _d+1 : [a, b] R. In practice it is an important question whether a certain regression function f _d+1, say, does or does not belong to the model. Therefore, we investigate the test problem H ₀ : "f _d+1 does not belong to the model" against K : "f _d+1 belong to the model" based on the least-squares residuals of the observations made at design points of the experimental region [a, b]. By a new functional central limit theorem given in Bischoff (1998, Ann. Statist. 26, 1398–1410), we are able to determine optimal tests in an asymptotic way. Moreover, we introduce the problem of experimental design for the optimal test statistics. Further, we compare the asymptotically optimal test with the likelihood ratio test (F-test) under the assumption that the error is normally distributed. Finally, we consider real change-point problems as examples and investigate by simulations the behavior of the asymptotic test for finite sample sizes. We determine optimal designs for these examples.     

On convergence subsystems of orthonormal systems

It is proved that for any sequence {R _k} _k=1 of real numbers satisfyingR _kk (k1) andR _k=o(k log₂ k),k, there exists an orthonormal system {n _k(x)} _n=1 ,x (0;1), such that none of its subsystems {n _k(x)} _k=1 withn _kR_k (k1) is a convergence subsystem.     

On conjugacy of some systems of functions

Summary. We investigate the bounded solutions j:[0,1]? X \varphi:[0,1]\to X of the system of functional equations¶¶j(f_k(x))=F_k(j(x)),    k=0,?,n-1,x ? [0,1] \varphi(f_k(x))=F_k(\varphi(x)),\;\;k=0,\ldots,n-1,x\in[0,1] ,(*)¶where X is a complete metric space, f₀,?,f_n-1:[0,1]?[0,1] f_0,\ldots,f_{n-1}:[0,1]\to[0,1] and F₀,...,F_n-1:X? X F_0,...,F_{n-1}:X\to X are continuous functions fulfilling the boundary conditions f₀(0) = 0, f_n-1(1) = 1, f_k+1(0) = f_k(1), F₀(a) = a,F_n-1(b) = b,F_k+1(a) = F_k(b), k = 0,?,n-2 f_{0}(0) = 0, f_{n-1}(1) = 1, f_{k+1}(0) = f_{k}(1), F_{0}(a) = a,F_{n-1}(b) = b,F_{k+1}(a) = F_{k}(b),\,k = 0,\ldots,n-2 , for some a,b ? X a,b\in X . We give assumptions on the functions f_k and F_k which imply the existence, uniqueness and continuity of bounded solutions of the system (*). In the case X = \Bbb C X= \Bbb C we consider some particular systems (*) of which the solutions determine some peculiar curves generating some fractals. If X is a closed interval we give a collection of conditions which imply respectively the existence of homeomorphic solutions, singular solutions and a.e. nondifferentiable solutions of (*).     

The summation of some series with variable coefficients by approximate analytical expressions

In a series of the form _k ^=1/ a _k f _k (x),a _k , which need be given only numerically, is approximated by an analytical functiong(k). If then _k ^=1/ g(k)f _k (x) can be summed exactly, it yields an approximation to the given series.Research sponsored by the U.S. Atomic Energy Commission under contract with the Union Carbide Corporation.     

An implicit function theorem: Comment

In Ref. 1, Jittorntrum proposed an implicit function theorem for a continuous mappingF:R ⁿ ×R ^m R ⁿ, withF(x ⁰,y ⁰)=0, that requires neither differentiability ofF nor nonsingularity of _x F(x ⁰,y ⁰). In the proof, the local one-to-one condition forF(·,y):A R ⁿ R ⁿ for ally B is consciously or unconsciously treated as implying thatF(·,y) mapsA one-to-one ontoF(A, y) for ally B, and the proof is not perfect. A proof can be given directly, and the theorem is shown to be the strongest, in the sense that the condition is truly if and only if.     

On some functional equations of Goląb-Schinzel type

Summary LetE be a real Hausdorff topological vector space. We consider the following binary law * on ·E:(, ) * (, ) = (, ^k + ) for(, ), (, ) × E where is a nonnegative real number,k andl are integers.In order to find all subgroupoids of ( ·E, *) which depend faithfully on a set of parameters, we have to solve the following functional equation:f(f(y) ^k x +f(x) ^l y) =f(x)f(y) (x, y E). (1)In this paper, all solutionsf: of (1) which are in the Baire class I and have the Darboux property are obtained. We obtain also all continuous solutionsf: E of (1). The subgroupoids of (^* ·E, *) which dapend faithfully and continuously on a set of parameters are then determined in different cases. We also deduce from this that the only subsemigroup ofL _n ¹ of the form {(F(x ₂,x ₃, ,x _n ),x ₂,x ₃, ,x _n ); (x ₂, ,x _n ) ^{n – 1} }, where the mappingF: ^{n – 1 *} has some regularity property, is {1} × ^{n – 1} .We may noitice that the Gob-Schinzel functional equation is a particular case of equation (1)(k = 0, l = 1, = 1). So we can say that (1) is of Gob—Schinzel type. More generally, whenE is a real algebra, we shall say that a functional equation is of Gob—Schinzel type if it is of the form:f(f(y) ^k x +f(x) ^l y) =F(x,y,f(x),f(y),f(xy)) wherek andl are integers andF is a given function in five variables. In this category of functional equations, we study here the equation:f(f(y) ^k x +f(x) ^l y) =f(xy) (x, y f: ). (4)This paper extends the results obtained by N. Brillouët and J. Dhombres in [3] and completes some results obtained by P. Urban in his Ph.D. thesis [11] (this work has not yet been published).Dedicated to the memory of Alexander M. Ostrowski on the occasion of the 100th anniversary of his birth     

Witten deformation of analytic torsion and the spectral sequence of a filtration

LetF be a flat vector bundle over a compact Riemannian manifoldM and letf :M be a Morse function. Letg ^F be a smooth Euclidean metric onF, letg _t ^F =e ^–2tf g ^F , and let ^RS (t) be the Ray-Singer analytic torsion ofF associated with the metricg _t ^F . Assuming that f satisfies the Morse-Smale transversality conditions, we provide an asymptotic expansion for log ^RS (t) fort+ of the forma ₀+a ₁ t+blog(t/)+o(1), where the coefficientb is a half-integer depending only on the Betti numbers ofF. In the case where all the critical values off are rational, we calculate the coefficientsa ₀ anda ₁ explicitly in terms of the spectral sequence of a filtration associated with the Morse function. These results are obtained as applications of a theorem by Bismut and Zhang.The research was supported by grant No. 449/94-1 from the Israel Academy of Sciences and Humanities.