Wiki
Wiki

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

Updated


Source. J. Koizumi, Irrationality of the reciprocal sum of doubly exponential sequences, arXiv:2504.05933v1 (8 April 2025); Remark 22 on p. 13, Example 11 on p. 8. Published as Integers 26 (2026), paper A28, where they are Remark 4 (p. 16) and Example 1 (p. 10). The editions are identified on the source card.

Read depth. Claims checked: the remark and Example 11 were read clause by clause on the page images of both editions. The transcendence of c(m)c(m) is the paper's citation of Dubickas and was not checked here; nothing here is independently reviewed.

Statement

For a positive integer mm, Example 11 (p. 8) defines s1(m)=m+1s_1(m)=m+1 and sn+1(m)=sn(m)2−sn(m)+1s_{n+1}(m)=s_n(m)^2-s_n(m)+1, the pseudo-greedy expansion of 1/m1/m, and cites the constant c(m)>1c(m)>1 with sn(m)≈c(m)2ns_n(m)\approx c(m)^{2^n} (ratio tending to 11), shown irrational by Wagner and Ziegler and transcendental by Dubickas.

Remark 22 (p. 13). Suppose Question 5 has an affirmative answer, equivalently that Conjecture 6 holds. Then the exceptional set of Theorem 4 is

(1,∞)∖I={c(m)2−N:N≥0, m>0},(1,\infty)\setminus\mathcal I=\{c(m)^{2^{-N}}: N\ge0,\ m>0\},

so every real algebraic α>1\alpha>1 lies in I\mathcal I; in particular 22n2^{2^n} would be a Type 2 irrationality sequence.

Proof pointer

Page 13. If an≈α2na_n\approx\alpha^{2^n} has rational reciprocal sum, the assumed answer makes the sequence eventually follow an+1=an2−an+1a_{n+1}=a_n^2-a_n+1, so aN+n=sn(m)a_{N+n}=s_n(m) for some N≥0N\ge0 and m>0m>0, and α=c(m)2−N\alpha=c(m)^{2^{-N}}. The transcendence of c(m)c(m) makes every such α\alpha transcendental.

Dependencies

Theorem 16 and Theorem 4 of the same paper; A. Dubickas, Ramanujan J. 57 (2022), 569--581, for the transcendence of c(m)c(m), as the paper cites it (not held here).

Bears on

  • Problem 263: a conditional implication. An affirmative answer to the question of Problem 243 would answer the first question of Problem 263 affirmatively; the remark proves nothing unconditionally, and says nothing on the second question.
  • Problem 243: the remark records a consequence of an affirmative answer, not progress on the question.