Wiki
Wiki

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

Updated


Source. Theorem 3.2, preprint p. 11, its proof on p. 12; Proposition 3.2, p. 9, with its proof on pp. 9--11. Read on the rendered pages. The edition read is identified on the source card.

Statement

Let (an)n≥1(a_n)_{n\ge1} be a sequence of positive integers such that aN,aN+1,…,a4Na_N,a_{N+1},\ldots,a_{4N} form a geometric sequence for infinitely many NN, and assume that

an=o(nn/7)(14)a_n=o\bigl(n^{n/7}\bigr)\qquad(14)

for nn sufficiently large. Then S=∑n=1∞an/n!∉QS=\sum_{n=1}^{\infty}a_n/n!\notin\mathbb{Q}.

Proposition 3.2 (p. 9)

The theorem is derived from the proof of this proposition, with a=1a=1, b=0b=0 and positive terms. Let a>0a>0 and b≥0b\ge0 be fixed integers and (an)(a_n) a sequence of Gaussian integers such that aN,…,a4Na_N,\ldots,a_{4N} form a geometric sequence for infinitely many NN, with an=o(nn/7)a_n=o(n^{n/7}) for nn sufficiently large (9). Then either S=∑n≥1an/(a+b)a,n∉Q[i]S=\sum_{n\ge1}a_n/(a+b)_{a,n}\notin\mathbb{Q}[i], or

aN∑n=0N−1a2N+n(N+n)!n! (2aN+b)a,N+n+1=o(N−N/8)a^N\sum_{n=0}^{N-1}a_{2N+n}\frac{(N+n)!}{n!\,(2aN+b)_{a,N+n+1}}=o\bigl(N^{-N/8}\bigr)

for infinitely many such NN. Here (x)a,n=x(x+a)⋯(x+(n−1)a)(x)_{a,n}=x(x+a)\cdots(x+(n-1)a).

Proof pointer

Proposition 3.2 (pp. 9--11) assumes S=t/qS=t/q and forms, for large NN with a geometric run, a Gaussian integer DND_N from the tails with NN-th difference weights. Lemma 2.3 reduces DND_N to the displayed main term plus o(N−N/8)o(N^{-N/8}), which bounds ∣DN∣|D_N| by (N!)6/7(N!)^{6/7} (12), while (13) makes DND_N divisible by N!/ANN!/A_N with N!/AN>(N!)6/7N!/A_N>(N!)^{6/7}; so DN=0D_N=0. For Theorem 3.2 (a=1a=1, b=0b=0, positive terms) every term of the main sum is positive, and Stirling's formula gives ∣DN∣>7−N|D_N|>7^{-N} for large NN (p. 12), so DN≠0D_N\ne0.

Dependencies

Proposition 3.2, Lemma 2.1 and Lemma 2.3 of the same paper. Theorem 4.2 applies this theorem: Theorem 4.2.

Bears on

No catalog problem directly.