dimensions of subsets of cantor-type sets维度cantor-type集的子集.pdfVIP

DIMENSIONS OF SUBSETS OF CANTOR-TYPE SETS FANG-XIONG ZHEN Received 15 March 2006; Accepted 9 July 2006 We define the Cantor-type set E first, and then the Besicovitch subset Bp of E. We mainly show the dimensions of subsets of Cantor-type set E in compatible case and incompatible case. Copyright © 2006 Hindawi Publishing Corporation. All rights reserved. 1. Introduction Let I = [0,1] be the unit interval on the real line and m 1 be an integer. LetJ = {0,1, ... , m − 1}. For every point x ∈ I , there is a unique base-m representation x = Σ∞ j k m−k k =1 with j k ∈ J except for countable many points. Since countable sets do not interfere with our work, we neglect them here. { For each j ∈ J , x ∈ [0,1], and n ∈ N, let τ (x, n) = k : i = j , 1 ≤ k ≤ n}, then the j k limit τ (x ) = lim (1/n)τ (x, n) is called the frequency of number j in the base-m repre- j n→∞ j sentation of x . Here and in the following context, the notation “A” denotes the number of elements in set A. A classical result of Borel [3] says that for Lebesgue almost every x ∈ [0,1], we have τj (x ) = 1/m. As for another problem, for a given probability vector p = (p 0 ,p 1, ... ,p m−1) such that Σj ∈J pj = 1, consider the set Λ p ,p , ... ,p = x ∈ [0,1] : τ (x ) = p , for j ∈ J . (1.1) m 0 1 m−1 j j That is, the set Λm (p 0 ,p 1, ... ,p m−1) is composed of the number in [0,1] having a ratio pj of digits equal to j in its base-m representation for each j . A precursor theory is due to