On the Largest Prime factor of the k k -generalized Pell numbers

Herbert Batte · 2025

Read On the Largest Prime factor of the k k -generalized Pell numbers by Herbert Batte (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-19052.txt.

Introduction

Herbert Batte, a solo author, studies the largest prime factor of generalized Pell numbers, in a 2025 math paper. The Pell sequence is a classic number sequence, and the k-generalized version sums the previous k terms with weights. Batte asks how large the biggest prime divisor must be as the sequence grows.

He proves that, from the fourth term on, the largest prime factor always exceeds a small fraction of the double logarithm of the index. He also finds every generalized Pell number whose largest prime factor is at most seven. The work extends prime divisor results known for simpler sequences.

Catalog record

CREATORHerbert Batte
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE42 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-19052

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

On the Largest Prime factor of the k k -generalized Pell numbers Herbert Batte Abstract Let k ≥ 2 k\geq 2 be an integer and consider the k k -generalized Pell sequence { P n ( k ) } n ≥ 2 − k \{P_{n}^{(k)}\}_{n\geq 2-k} , defined by the initial values 0 , … , 0 , 0 , 1 0,\ldots,0,0,1 (a total of k k terms), and the recurrence P n ( k ) = 2 ​ P n − 1 ( k ) + P n − 2 ( k ) + ⋯ + P n − k ( k ) P_{n}^{(k)}=2P_{n-1}^{(k)}+P_{n-2}^{(k)}+\cdots+P_{n-k}^{(k)} , for all n ≥ 2 n\geq 2 . For any integer m m , let 𝒫 ⁡ ( m ) \mathcal{P}(m) denote the largest prime factor of m m , with the convention 𝒫 ⁡ ( 0 ) = 𝒫 ⁡ ( ± 1 ) = 1 \mathcal{P}(0)=\mathcal{P}(\pm 1)=1 . In this paper, we prove that for n ≥ 4 n\geq 4 , the inequality 𝒫 ⁡ ( P n ( k ) ) > ( 1 / 104 ) ​ log ⁡ log ⁡ n \mathcal{P}(P_{n}^{(k)})>(1/104)\log\log n holds. Additionally, we find all k k -generalized Pell numbers P n ( k ) P_{n}^{(k)} , whose largest prime factor does not exceed 7 7 . Keywords and phrases : k -generalized Pell numbers; greatest prime factor; linear forms in logarithms; LLL-reduction. 2020 Mathematics Subject Classification : 11B39, 11D61, 11D45, 11Y50. ∗ Corresponding author 1 Introduction 1.1 Background Let k\geq 2 be a fixed integer. The sequence of k -generalized Pell numbers, denoted by \{P_{n}^{(k)}\} , is defined recursively by P_{n}^{(k)}=2P_{n-1}^{(k)}+P_{n-2}^{(k)}+\cdots+P_{n-k}^{(k)},\qquad\text{for all }\qquad n\geq 2, with initial terms given by P_{2-k}^{(k)}=\cdots=P_{-1}^{(k)}=P_{0}^{(k)}=0 and P_{1}^{(k)}=1 . When k=2 , the sequence coincides with the classical Pell numbers, and in that case we drop the superscript (k) for simplicity. The first few values for P_{n}^{(k)} , for some small values of k , are given in Table 1.1 below. Table 1.1: First non-zero k -Pell numbers k Name First non-zero terms 2 Pell 1, 2, 5, 12, 29, 70, 169, 408, 985, 2378, 5741, 13860, 33461, \ldots 3 3-Pell 1, 2, 5, 13, 33, 84, 214, 545, 1388, 3535, 9003, 22929, 58396, \ldots 4 4-Pell 1, 2, 5, 13, 34, 88, 228, 591, 1532, 3971, 10293, 26680, 69156, \ldots 5 5-Pell 1, 2, 5, 13, 34, 89, 232, 605, 1578, 4116, 10736, 28003, 73041, \ldots 6 6-Pell 1, 2, 5, 13, 34, 89, 233, 609, 1592, 4162, 10881, 28447, 74371, \ldots 7 7-Pell 1, 2, 5, 13, 34, 89, 233, 610, 1596, 4176, 10927, 28592, 74815, \ldots 8 8-Pell 1, 2, 5, 13, 34, 89, 233, 610, 1597, 4180, 10941, 28638, 74960, \ldots 9 9-Pell 1, 2, 5, 13, 34, 89, 233, 610, 1597, 4181, 10945, 28652, 75006, \ldots 10 10-Pell 1, 2, 5, 13, 34, 89, 233, 610, 1597, 4181, 10946, 28656, 75020, \ldots For any integer m , we denote by \mathcal{P}(m) the largest prime divisor of m , adopting the convention that \mathcal{P}(0)=\mathcal{P}(\pm 1)=1 . Investigating lower bounds for the largest prime factor in linear recurrence sequences is a problem that has attracted considerable attention in Number theory. Some important contributions have been made in this area (see, for instance, [ 2 ] and [ 5 ] ). In this paper, we examine the properties of the k -generalized Pell numbers with the objective of establishing explicit lower bounds for \mathcal{P}(P_{n}^{(k)}) in terms of the parameters k and n . Note that since P_{1}^{(k)}=1 , P_{2}^{(k)}=2 and P_{3}^{(k)}=5 for all k\geq 2 , we have that \mathcal{P}(P_{n}^{(k)}) is either 1, 2 or 5, for all k\geq 2 and n\leq 3 . Therefore, we work with n\geq 4 . We prove the following results. 1.2 Main Results Theorem 1.1 . Let \{P_{n}^{(k)}\}_{n\geq 2-k} be the sequence of k -generalized Pell numbers. Then, the inequality \displaystyle\mathcal{P}(P_{n}^{(k)})>\dfrac{1}{104}\log\log n, holds for all n\geq 4 and k\geq 2 . Theorem 1.2 . The only solutions to the Diophantine equation \displaystyle P_{n}^{(k)}=2^{a}\cdot 3^{b}\cdot 5^{c}\cdot 7^{d}, (1.1) in non-negative integers n , k , a , b , c , d , e with k\geq 2 and n\geq 4 , are \displaystyle P_{4}^{(2)}=12,\qquad P_{6}^{(2)}=70,\qquad P_{6}^{(3)}=84,\qquad\text{and}\qquad P_{10}^{(5)}=4116. Our approach can be summarized as follows. (a) We utilize lower bounds for linear forms in logarithms of algebraic numbers to derive an upper estimate for \log n in terms of the parameters k and \mathcal{P}(P_{n}^{(k)}) . The analysis is more straightforward when k is small. For larger values of k , we take advantage of the fact that the dominant root of the k -generalized Pell sequence is very close to \varphi^{2} , where \varphi:=\dfrac{1+\sqrt{5}}{2}, is the golden ratio. This allows us to simplify our computations by approximating this root with \varphi in the relevant logarithmic expressions. (b) Moreover, the techniques described here, combined with the use of the LLL-algorithm, are applied in the concluding part of the paper to determine all k -generalized Pell numbers whose largest prime factor is at most 7 . 2 Methods 2.1 Some identities on k -Pell numbers We begin with an identity from [ 9 ] , which relates P_{n}^{(k)} with the m -th Fibonacci number. It states that \displaystyle P_{n}^{(k)}=F_{2n-1}, for all 1\leq n\leq k+1 , while the next term is P_{k+2}^{(k)}=F_{2k+3}-1 . The characteristic polynomial of \{P_{n}^{(k)}\}_{n\in{\mathbb{Z}}} is given by \Psi_{k}(x)=x^{k}-2x^{k-1}-\cdots-x-1. This polynomial is irreducible in \mathbb{Q}[x] and it possesses a unique real root \alpha:=\alpha(k)>1 . All the other roots of \Psi_{k}(x) are inside the unit circle, see [ 4 ] . The particular root \alpha can be found in the interval \displaystyle\varphi^{2}(1-\varphi^{-k})<\alpha<\varphi^{2},\qquad\text{for}\qquad k\geq 2, (2.1) as noted in [ 4 ] . As in the classical case when k=2 , it was shown in [ 4 ] that \displaystyle\alpha^{n-2}\leq P_{n}^{(k)}\leq\alpha^{n-1},\qquad\text{holds for all}\qquad n\geq 1\quad\text{and}\quad k\geq 2. (2.2) Next, for k\geq 2 , we define \displaystyle f_{k}(x):=\dfrac{x-1}{(k+1)x^{2}-3kx+k-1}=\dfrac{x-1}{k(x^{2}-3x+1)+x^{2}-1}. (2.3) In Lemma 1 of [ 3 ] , Bravo and Herera showed that \displaystyle 0.276<f_{k}(\alpha)<0.5\qquad\text{and}\qquad|f_{k}(\alpha_{i})|<1, (2.4) holds for all 2\leq i\leq k , where \alpha_{i} for i=2,\ldots,k are the roots of \Psi_{k}(x) inside the unit circle. So, the number f_{k}(\alpha) is not an algebraic integer. Indeed, if we suppose that f_{k}(\alpha) is an algebraic integer, then we know that f_{k}(\alpha)\neq 0 and N_{\mathbb{Q}(\alpha)/\mathbb{Q}}(f_{k}(\alpha))\in\mathbb{Z} , where N_{\mathbb{Q}(\alpha)/\mathbb{Q}}(f_{k}(\alpha)) is the norm of f_{k}(\alpha) from \mathbb{Q}(\alpha) to \mathbb{Q} . Therefore, we obtain 1\leq\left|N_{\mathbb{Q}(\alpha)/\mathbb{Q}}(f_{k}(\alpha))\right|=f_{k}(\alpha)\cdot\prod_{i=2}^{k}\left|f_{k}(\alpha^{(i)})\right|<0.5, which is absurd. So f_{k}(\alpha) is not an algebraic integer. Moreover, it was also shown in Theorem 3.1 of [ 4 ] that \displaystyle P_{n}^{(k)}=\displaystyle\sum_{i=1}^{k}f_{k}(\alpha_{i})\alpha_{i}^{n}\qquad\text{and}\qquad\left|P_{n}^{(k)}-f_{k}(\alpha)\alpha^{n}\right|<\dfrac{1}{2}, (2.5) hold for all k\geq 2 and n\geq 2-k . The first part of ( 2.5 ) provides a Binet-type representation for P_{n}^{(k)} . In addition, the second inequality in ( 2.5 ) indicates that the impact of the roots located inside the unit circle on P_{n}^{(k)} is small. Next, let n\geq 4 and \displaystyle P_{n}^{(k)}=p_{1}^{\beta_{1}}\cdot p_{2}^{\beta_{2}}\cdots p_{s}^{\beta_{s}}, (2.6) be the prime factorization of the positive integer P_{n}^{(k)} , where 2=p_{1}<p_{2}<\cdots<p_{s} is the increasing sequence of prime numbers and the integers \beta_{i} , for i=1,2,\ldots,s , are non-negative integers. By the right hand side of relation ( 2.2 ), we have \log P_{n}^{(k)}\leq(n-1)\log\alpha<2(n-1)\log\varphi, since \alpha<\varphi^{2} , for all k\geq 2 in ( 2.1

Read the complete text