On the Box Dimension of Typical Measures

 Let be a complete metric space and let be the space of all probability Borel measures on X. We give some estimations of the upper and lower box dimensions of the typical (in the sense of Baire category) measure in . Received 29 November 2000; in final form 8 January 2002     

Relationships between Different Dimensions of a Measure

 We establish various relationships of the Hausdorff dimension, entropy dimension and -dimension of a measure without assuming that the local dimension of μ exists μ-a.e. These extend a well known result of Young. Received 2 October 2000; in revised form 20 September 2001     

The Dimension of Subsets of Moran Sets Determined by the Success Run Behaviour of Their Codings

 Let be a Moran set associated with the set . Let Γ be a non-empty subset of with non-empty complement. Associated with the behaviour of success run of symbols from Γ in the coding space is a decomposition of F such that

Intersections of Sets Starshaped Via Paths of Bounded Length

 Let S be a nonempty closed, simply connected set in the plane. For α > 0, let ℳ denote the family of all maximal subsets of S which are starshaped via paths of length at most α. Then ⋂{M : M in ℳ} is either starshaped via α-paths or empty. The result fails without the simple connectedness condition. However, even with a simple connectedness requirement, there is no Helly theorem for intersections of sets which are starshaped via α-paths. Received November 19, 2001; in revised form April 25, 2002 Published online November 18, 2002     

Small Sets of Divergence Points are Dimensionless

Let be a subshift modelled by a strongly connected graph, and let denote the shift. For n , let be the nth empirical measure, i.e.

On the Curvature of the Free Boundary for the Obstacle Problem in Two Dimensions

We give a new proof of the fact that the free boundary for the obstacle problem in two dimensions satisfies a natural and sharp inner ball condition.     

Porosity and diametrically maximal sets in C(K)

This paper is motivated by recent attempts to investigate classical notions from finite-dimensional convex geometry in spaces of continuous functions. Let be the family of all closed, convex and bounded subsets of C(K) endowed with the Hausdorff metric. A completion of is a diametrically maximal set satisfying A ⊂ D and diam A = diam D. Using properties of semicontinuous functions and an earlier result by Papini, Phelps and the author [12], we characterize the family γ(A) of all possible completions of . We give also a formula which calculates diam γ(A) and prove finally that, if K is a Hausdorff compact space with card K > 1, then the family of those elements of having a unique completion is uniformly very porous in with a constant of lower porosity greater than or equal to 1/3.     

Measures on the Product of Compact Spaces

 If K is an uncountable metrizable compact space, we prove a “factorization” result for a wide variety of vector valued Borel measures μ defined on K ⁿ . This result essentially says that for every such measure μ there exists a measure μ′ defined on K such that the measure μ of a product A ₁ × ⋯ × A _n of Borel sets of K equals the measure μ′ of the intersection A ₁′∩⋯∩A _n ′, where the A _i ′’s are certain transforms of the A _i ’s. Partially supported by DGICYT grant PB97-0240. Received August 23, 2001; in revised form March 21, 2002     

Packing of Incongruent Circles on the Sphere

 In this paper we provide an upper bound to the density of a packing of circles on the sphere, with radii selected from a given finite set. This bound is precise, e.g. for the system of incircles of Archimedean tilings (4, 4, n) with n ? 6. A generalisation to the weighted density of packing is applied to problems of solidity of a packing of circles. The simple concept of solidity was introduced by L. Fejes Toóth [6]. In particular, it is proved that the incircles of the faces of the Archimedean tilings (4,6,6), (4,6,8) and (4, 6, 10) form solid packings. (Received 21 August 2000; in revised form 21 March 2001)     

Sets of Locally Maximal Visibility and Finite Unions of Starshaped Sets

 Let , where is an open connected subset of some linear topological space, such that S contains all triangular regions whose (relative) boundaries lie in S. If some finite subset T of S has locally maximal visibility in S, then . Hence S is a finite union of starshaped sets whose kernels are determined by T. An analogous result holds for S open. Moreover, counterexamples show that neither the requirement on triangular regions nor the restriction to a finite set T can be deleted. (Received 7 September 1998; in revised form 25 October 1999)     

Disasters in topology without the axiom of choice

We show that some well known theorems in topology may not be true without the axiom of choice. Received: 29 August 1995 / Revised version: 23 June 2000 / Published online: 3 October 2001     

The Singularity Spectrum f(α) of Some Moran Fractals

We show that the multifractal decomposition behaves as expected for a family of sets E known as homogeneous Moran fractals associated with the Fibonacci sequence , using probability measures () associated with the Fibonacci sequence . For each value of a parameter (_min, _max), we define multifractal components E of E, and show that they are fractals in the sense of Taylor. We give the explicit formula for the dimension of E. Also our method can be used for the Moran fractals associated with some more general sequences.     

On the false pole problem

We prove that if a convex body has an interior false pole with respect to some hyperplane, then the body is an ellipsoid. This research was partially carried out during the postdoctoral visit of this author at University College London, and it was supported by CONACYT, México.     

Embedding metric spaces into normed spaces and estimates of metric capacity

Let be an arbitrary real normed space of finite dimension d ≥ 2. We define the metric capacity of as the maximal such that every m-point metric space is isometric to some subset of (with metric induced by ). We obtain that the metric capacity of lies in the range from 3 to , where the lower bound is sharp for all d, and the upper bound is shown to be sharp for d ∈ {2, 3}. Thus, the unknown sharp upper bound is asymptotically linear, since it lies in the range from d + 2 to . Research supported by the German Research Foundation, Project AV 85/1-1.     

Empirical multifractal moment measures of self-similar measure for <Emphasis Type="Italic">q</Emphasis> < 0<Superscript>*</Superscript>

For q ≥ 0, Olsen [1] has attained the exact rate of convergence of the L ^q -spectrum of a self-similar measure and showed that the so-called empirical multifractal moment measures converges weakly to the normalized multifractal measures. Unfortunately, nothing is known for q < 0. Indeed, the problem of analysing the L ^q - spectrum for q < 0 is generally considered significantly more difficult since the L ^q -spectrum is extremely sensitive to small variations of μ for q < 0. In [2] we showed that self-similar measures satisfying the Open Set Condition (OSC) are Ahlfors regular and, using this fact, we obtained the exact rate of convergence of the L ^q -spectrum of a self-similar measure satisfying the OSC for q < 0. In this paper, we apply the results from [2] to show the empirical multifractal q’th moment measures of self-similar measures satisfying the OSC converges weakly to the normalized multifractal Hausdorff measures for q < 0. Authors’ addresses: Jiaqing Xiao, School of Science, Wuhan University of Technology, Wuhan 430070, China; Wu Min, School of Mathematical Sciences, South China University of Technology, Guangzhou, 510640, China     

Topological dimensions and multifractal Rényi dimensions of Polish spaces: a multifractal Szpilrajn type theorem

In this paper we consider the relationship between the topological dimension and the lower and upper q-Rényi dimensions and of a Polish space X for q ∈ [1, ∞]. Let and denote the Hausdorff dimension and the packing dimension, respectively. We prove that for all analytic metric spaces X (whose upper box dimension is finite) and all q ∈ (1, ∞); of course, trivially, for all q ∈ [1, ∞]. As a corollary to this we obtain the following result relating the topological dimension and the lower and upper q-Rényi dimensions: for all Polish spaces X and all q ∈ [1, ∞]; in (1) and (2) we have used the following notation, namely, for two metric spaces X and Y, we write X ∼ Y if and only if X is homeomorphic to Y. Equality (1) has recently been proved for q = ∞ by Myjak et al. Author’s address: Department of Mathematics, University of St. Andrews, St. Andrews, Fife KY16 9SS, Scotland     

Optimizing Area and Perimeter of Convex Sets for Fixed Circumradius and Inradius

 In this paper two problems posed by Santaló are solved: we determine the planar convex sets which have maximum and minimum area or perimeter when the circumradius and the inradius are given, obtaining complete systems of inequalities for the cases (A, R, r) and (p, R, r). This work is supported in part by Dirección General de Investigación (MCYT) BFM2001-2871, and by OTKA grants No 31984 and 30012 Received October 15, 2001; revised January 29, 2002     

Simultaneous Packing and Covering in the Euclidean Plane

 In this article we study the simultaneous packing and covering constants of two-dimensional centrally symmetric convex domains. Besides an identity result between translative case and lattice case and a general upper bound, exact values for some special domains are determined. Similar to Mahler and Reinhardt’s result about packing densities, we show that the simultaneous packing and covering constant of an octagon is larger than that of a circle. (Received 17 January 2001; in revised form 13 July 2001)     

Egoroff’s theorem and maximal run length

We construct a sequence of measurable functions converging at each point of the unit interval, but the set of points with any given rate of convergence has Hausdorff dimension one. This is used to show that a version of Egoroff’s theorem due to Taylor is best possible. The construction relies on an analysis of the maximal run length of ones in the dyadic expansion of real numbers. It is also proved that the exceptional set for a limit theorem of Renyi has Hausdorff dimension one.