Wiki
Wiki

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

Updated


Source. Theorem 2.2, p. 4, of Mikhail Neklyudov, Functional analysis approach to the Collatz conjecture, arXiv:2106.11859v9 (2022), published in Results Math. 79 (2024), no. 4, Paper No. 140, in the edition identified on the source card.

Statement

TT is the reduced Collatz map on Z\mathbb Z and T\mathcal T the operator with T(zn)=zT(n)\mathcal T(z^n)=z^{T(n)}, as on the Lemma 1.1 page. In Section 2 (p. 3) T\mathcal T acts on the quotient Hber2(D)/XH^2_{ber}(D)/X of the Bergman space of the unit disc, where X=span⁡{1,z,z2}X=\operatorname{span}\{1,z,z^2\} is invariant, and has norm at most 22 there.

Theorem 2.2 (p. 4). Quoted: "If TT has no nontrivial cycles than [sic] T\mathcal T is hypercyclic."

Hypercyclic means that some vector has a dense orbit under T\mathcal T. The operator acts on power series, so only the values of TT on the nonnegative integers enter, and the cycles meant are cycles in N\mathbb N; on the negative integers TT has other cycles, such as the fixed point −1-1.

Read depth. Claims checked: the statement was read clause by clause on p. 4; the proof was read through, not checked step by step.

Proof pointer

Page 4. The proof writes out only the case in which the 3n+13n+1 conjecture holds and says that the case of a diverging trajectory is similar. It applies the Godefroy--Shapiro hypercyclicity criterion (their Corollary 1.5, p. 235) with the right inverse Sg(z)=g(z2)Sg(z)=g(z^2). The dense set on which iterates of T\mathcal T vanish comes from Lemma 2.1 (p. 3), which under the conjecture gives Hber2(D)/X=⋃n≥1Ker⁡(Tn)‾H^2_{ber}(D)/X=\overline{\bigcup_{n\ge1}\operatorname{Ker}(\mathcal T^n)}; the iterates of SS tend to zero on monomials.

Dependencies

Lemma 2.1 of the same paper (p. 3), which also proves T\mathcal T surjective on Hber2(D)/XH^2_{ber}(D)/X.

Bears on

  • #1135: a consequence of the absence of nontrivial cycles, which a positive answer to the problem would imply; it proves no case of the problem.