Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated

Nguyen (2022): Sparse Fibonacci reciprocal series

../

theorem_1_2: Nguyen's transcendence theorem: if c > 2 and the positive integers n_1 < n_2 < ... satisfy n_{k+1}/n_k >= c for every k, then the sum of 1/f_k is transcendental for any choice of f_k among F_{n_k} and L_{n_k}.

theorem_1_3: Nguyen's main theorem: for a real quadratic unit alpha, rational terms u_n = a_n alpha^n - b_n beta^n with coefficients of subexponential height, and indices with n_{k+1}/n_k >= c > 2, the sum of c_{n_k}/u_{n_k} is transcendental.


Khoa Dang Nguyen, Transcendental series of reciprocals of Fibonacci and Lucas numbers, Algebra & Number Theory 16 (2022), no. 7, 1627--1654, doi:10.2140/ant.2022.16.1627. The result labels below were checked in arXiv:2009.02446v1, rather than the publisher's text.

Question 1.1 (p. 1) repeats the Erdős--Graham question in Problem 267, which asks whether ∑k1/Fnk\sum_k1/F_{n_k} is irrational whenever nk+1/nk≥c>1n_{k+1}/n_k\ge c>1. The paper's results:

  • Theorem 1.2 (p. 1): if c>2c>2 and n1<n2<⋯n_1<n_2<\cdots are positive integers with nk+1/nk≥cn_{k+1}/n_k\ge c for every kk, then
∑k=1∞1fk\sum_{k=1}^{\infty}\frac{1}{f_k}

is transcendental for any choice of fk∈{Fnk,Lnk}f_k\in\{F_{n_k},L_{n_k}\}, where LnL_n is the Lucas sequence: L1=1L_1=1 and Ln=Fn−1+Fn+1L_n=F_{n-1}+F_{n+1} for n≥2n\ge2. The print's definition reads Ln=Fn−1+FnL_n=F_{n-1}+F_n, a misprint, as Example 1.4's Ln=αn+βnL_n=\alpha^n+\beta^n shows.

  • Theorem 1.3 (p. 3), the main theorem: the same conclusion for ∑kcnk/unk\sum_kc_{n_k}/u_{n_k}, where α≠±1\alpha\ne\pm1 is a real quadratic unit with conjugate β\beta, ∣β∣<1<∣α∣|\beta|<1<|\alpha|; an,bn,cna_n,b_n,c_n are real with cn∈Qc_n\in\mathbb Q, an,bn∈Q(α)a_n,b_n\in\mathbb Q(\alpha) and un=anαn−bnβn∈Qu_n=a_n\alpha^n-b_n\beta^n\in\mathbb Q; their logarithmic Weil heights are o(n)o(n); and unk≠0u_{n_k}\ne0, cnk≠0c_{n_k}\ne0 for every kk. Theorem 1.2 is its special case (Example 1.4, p. 3). The proof (Sections 3--5, pp. 5--26) uses the Subspace Theorem.

The paper points out (p. 2) that irrationality for c>2c>2 already follows from the estimate Fn1⋯FnN=o(FnN+1)F_{n_1}\cdots F_{n_N}=o(F_{n_{N+1}}); the new conclusion is transcendence. The abstract states that the bound c>2c>2 is best possible, because of the identity ∑k≥01/F2k=(7−5)/2\sum_{k\ge0}1/F_{2^k}=(7-\sqrt5)/2 (the Millin series, p. 1), whose index ratios equal 22. Theorem 1.2 says nothing about sequences whose ratios are bounded below only by a constant c≤2c\le2.

Source: https://doi.org/10.2140/ant.2022.16.1627.

Edition read. The copy read for this card is the arXiv v1 PDF (arXiv:2009.02446v1). The arXiv record names arXiv's non-exclusive distribution license (arXiv:2009.02446), every other right reserved.

Bears on. #267: Theorem 1.2 with fk=Fnkf_k=F_{n_k} gives transcendence, hence irrationality, of ∑k1/Fnk\sum_k1/F_{n_k} when the ratios nk+1/nkn_{k+1}/n_k are at least a constant c>2c>2; it does not reach constants c≤2c\le2.

Read status. Claims checked: Question 1.1, Theorems 1.2 and 1.3, Examples 1.4 and 1.5, and the comparison with the elementary irrationality bound were read in arXiv v1, pp. 1--3. The proof was read for its structure only and was not checked.

No file of this source is held: no license on record permits its redistribution, and the card cites the edition it names above.