Wiki
Wiki

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

Updated

Lai 2021 largest prime divisor

../

least_prime_divisor_bound: States an improved upper bound for the odd-index liminf of the least prime divisor of n!+1.

lemma_2_7: Bounds a sum of running minima of p-adic valuations over distinct shifted factorial values.

theorem_1_1: Gives a 1+9 log 2 limsup bound, with a positive-lower-density strengthening, for every nonzero polynomial shift of n!.


Li Lai, On the largest prime divisor of n!+1n!+1, arXiv:2103.14894v1 (27 March 2021).

The selected arXiv v1 PDF has 11 physical pages. It is a factorial-sequence paper; it does not concern the exponential sequence 2n−12^n-1 of Problem 977 and is not direct progress on that problem. The arXiv record names arXiv's non-exclusive distribution license (arXiv:2103.14894), every other right reserved.

For every nonzero f∈Z[X]f\in\mathbb Z[X], Theorem 1.1 proves

lim sup⁡n→∞P(n!+f(n))n≥1+9log⁡2≈7.238.\limsup_{n\to\infty}\frac{P(n!+f(n))}{n}\geq1+9\log2\approx7.238.

More strongly, for each fixed ε>0\varepsilon>0 the inequality P(n!+f(n))>(1+9log⁡2−ε)nP(n!+f(n))>(1+9\log2-\varepsilon)n, together with n!+f(n)>1n!+f(n)>1, holds on a set of positive integers nn whose lower asymptotic density is positive. This improves the constants 5/25/2 of Luca--Shparlinski for general polynomial shifts and 11/211/2 of Stewart for f=1f=1.

The new ingredient is Lemma 2.7, a simultaneous pp-adic valuation bound over an ordered family of distinct factorial arguments. The proof of Theorem 1.1 in Section 3 applies that estimate to the prime-power contributions in a large product of values n!+f(n)n!+f(n). The introduction also states two further applications of the method: an improved upper bound, approximately 1.2931.293, for the odd-index liminf of the least prime divisor of n!+1n!+1, and, on the same p. 2, an improvement of Luca--Shparlinski's lower bound (2π2+3)/18≈1.263(2\pi^2+3)/18\approx1.263 for lim sup⁡n→∞P(n!+2n−1)/n\limsup_{n\to\infty}P(n!+2^n-1)/n to 1+2π2−156log⁡32≈1.3201+\frac{2\pi^2-15}{6}\log\frac32\approx1.320. The paper gives no separate derivation of either application.

Results.

The result pages contain exact statements and proof pointers only. No complete proof is transcribed.

Bears on. #977 (context only): Theorem 1.1 with f=1f=1 gives lim sup⁡n→∞P(n!+1)/n≥1+9log⁡2\limsup_{n\to\infty}P(n!+1)/n\geq1+9\log2 for the factorial variant that the problem's catalog remarks mention; it does not show that P(n!+1)/nP(n!+1)/n tends to infinity and says nothing about P(2n−1)/nP(2^n-1)/n.

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