Abstract:Let β > 1 be a non-integer. We consider β-expansions of the form ∞ i=1 d i /β i , where the digits (d i ) i≥1 are generated by means of a Borel map K β defined on {0, 1} N ×[0, β /(β − 1)]. We show that K β has a unique mixing measure ν β of maximal entropy with marginal measure an infinite convolution of Bernoulli measures. Furthermore, under the measure ν β the digits (d i ) i≥1 form a uniform Bernoulli process. In case 1 has a finite greedy expansion with positive coefficients, the measure of maximal entrop… Show more
“…Properties L1 and L2 are well known and have appeared in several articles and with quite short proofs, e.g. [6] and [1], where in the latter paper simple and short dynamical proofs are given. We give the above proofs for two reasons; first, they are elementary and second, to make this exposition self-contained.…”
Section: Lazy Expansionsmentioning
confidence: 99%
“…Let q ∈ (1,2]. By an expansion with respect to q, or q-expansion, of a positive real number x we mean a sequence (e i ) i≥1 ⊆ {0, 1} satisfying…”
Section: Introductionmentioning
confidence: 99%
“…In 1991, Erdös, Horváth, and Joo [3] showed that for almost all q ∈ (1, 2], there are uncountably many different q-expansions, and surprisingly, there exist as well uncountably many exceptional q ∈ (0, 1) for which there is only one q-expansion. In 1998, Komornik and Loreti [5] determined the smallest base q ∈ (1,2] for which the q-expansion of 1 is unique. In 1999, Komornik and Loreti [6] gave a sufficient condition for which the number 1 has exactly two different q-expansions as well as using this information to construct the smallest base q for which the number 1 has exactly two different q-expansions.…”
Section: Introductionmentioning
confidence: 99%
“…Let q ∈ (1,2]. By a q-expansion of 1, we mean a sequence (e i ) i≥1 of integers in {0, 1} satisfying the equality 1 =…”
Abstract. The works of Erdös et al. about expansions of 1 with respect to a non-integer base q, referred to as q-expansions, are investigated to determine how far they continue to hold when the number 1 is replaced by a positive number x. It is found that most results about q-expansions for real numbers greater than or equal to 1 are in somewhat opposite direction to those for real numbers less than or equal to 1. The situation when a real number has a unique q-expansion, and when it has exactly two q-expansions are studied. The smallest base number q yielding a unique q-expansion is determined and a particular sequence is shown, in certain sense, to be the smallest sequence whose corresponding base number q yields exactly two q-expansions.
“…Properties L1 and L2 are well known and have appeared in several articles and with quite short proofs, e.g. [6] and [1], where in the latter paper simple and short dynamical proofs are given. We give the above proofs for two reasons; first, they are elementary and second, to make this exposition self-contained.…”
Section: Lazy Expansionsmentioning
confidence: 99%
“…Let q ∈ (1,2]. By an expansion with respect to q, or q-expansion, of a positive real number x we mean a sequence (e i ) i≥1 ⊆ {0, 1} satisfying…”
Section: Introductionmentioning
confidence: 99%
“…In 1991, Erdös, Horváth, and Joo [3] showed that for almost all q ∈ (1, 2], there are uncountably many different q-expansions, and surprisingly, there exist as well uncountably many exceptional q ∈ (0, 1) for which there is only one q-expansion. In 1998, Komornik and Loreti [5] determined the smallest base q ∈ (1,2] for which the q-expansion of 1 is unique. In 1999, Komornik and Loreti [6] gave a sufficient condition for which the number 1 has exactly two different q-expansions as well as using this information to construct the smallest base q for which the number 1 has exactly two different q-expansions.…”
Section: Introductionmentioning
confidence: 99%
“…Let q ∈ (1,2]. By a q-expansion of 1, we mean a sequence (e i ) i≥1 of integers in {0, 1} satisfying the equality 1 =…”
Abstract. The works of Erdös et al. about expansions of 1 with respect to a non-integer base q, referred to as q-expansions, are investigated to determine how far they continue to hold when the number 1 is replaced by a positive number x. It is found that most results about q-expansions for real numbers greater than or equal to 1 are in somewhat opposite direction to those for real numbers less than or equal to 1. The situation when a real number has a unique q-expansion, and when it has exactly two q-expansions are studied. The smallest base number q yielding a unique q-expansion is determined and a particular sequence is shown, in certain sense, to be the smallest sequence whose corresponding base number q yields exactly two q-expansions.
“…A wellknown class of symbolic dynamical systems is that of the β-shifts introduced by Rényi [26], developed by Parry in the seminal paper [25], and studied intensively thereafter, see for example [15,28,37,21,5,10,29,30,8,7,1].…”
For every real number β > 1, the β-shift is a dynamical system describing iterations of the map x → βx mod 1 and is studied intensively in number theory. Each β-shift has an associated language of finite strings of characters; properties of this language are studied for the additional insight they give into the dynamics of the underlying system.We prove that the language of the β-shift is recursive iff β is a computable real number. That fact yields a precise characterization of the reals: The real numbers β for which we can compute arbitrarily good approximations-hence in particular the numbers for which we can compute their expansion to some base-are precisely those for which there exists a program that decides for any finite sequence of digits whether the sequence occurs as a subword of some element of the β-shift.While the "only if" part of the proof of the above result is constructive, the "if" part is not. We show that no constructive proof of the "if" part exists. Hence, there exists no algorithm that transforms a program computing arbitrarily good approximations of a real number β into a program deciding the language of the β-shift.
Let β > 1 and let m > β be an integer. Each x ∈ I β := [0, m−1 β−1 ] can be represented in the formwhere ε k ∈ {0, 1, . . . , m − 1} for all k (a β-expansion of x). It is known that a.e. x ∈ I β has a continuum of distinct β-expansions. In this paper we prove that if β is a Pisot number, then for a.e. x this continuum has one and the same growth rate. We also link this rate to the Lebesgue-generic local dimension for the Bernoulli convolution parametrized by β.When β < 1+√ 5 2 , we show that the set of β-expansions grows exponentially for every internal x.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.