Determine the extremal liminf ratio of maximal term to maximum modulus for entire functions (Erdős #513)
Statement
Let $f=\sum_{n=0}^\infty a_nz^n$ be a transcendental entire function. Write $\mu(r)=\max_n \lvert a_n r^n\rvert$ for the maximal term of the series at radius $r$, and $M(r)=\max_{\lvert z\rvert=r}\lvert f(z)\rvert$ for the maximum modulus. What is the greatest possible value of $$\liminf_{r\to \infty} \frac{\mu(r)}{M(r)}\,?$$ Equivalently, determine $B$, the supremum over all transcendental entire $f$ of $\liminf_{r\to\infty}\mu(r)/M(r)$.
Acceptance. FULLY RESOLVES: determine $B$ exactly — a construction (or limiting family) of transcendental entire functions attaining the value in the liminf, plus a matching upper-bound proof valid for all transcendental entire functions; a complete written proof with all steps, machine-checkable (Lean/Coq) preferred. ADVANCES: (a) an explicit entire function, given by its coefficient sequence, together with a rigorous (certified, e.g. interval-arithmetic) proof that its liminf ratio strictly exceeds the best lower bound stated in the background; (b) an upper bound strictly below the best upper bound stated in the background, with proof; or (c) an explicit numeric value for the constant $c$ in the Clunie–Hayman upper bound $2/\pi - c$, with proof. Deliver the proof file, or the construction (coefficient formula) plus certification code and its output.
Background
Posed by Erdős [Er61, p.249]; listed as open on erdosproblems.com/513 (fetched 2026-07-13, status 'open', tagged 'analysis'). Since $\mu(r)\le M(r)$ always, and classical examples give ratio $1/2$ in the liminf, it is trivial that $B\in[1/2,1]$. Kövári (unpublished) observed $B>1/2$. Gray and Shah [GrSh63] recorded an argument of Clunie giving $B\le 2/\pi\approx 0.63662$, and Clunie and Hayman [ClHa64] improved both sides to $4/7 < B \le 2/\pi - c$ for some absolute constant $c>0$ (note $4/7\approx 0.57143$). Recently He and Tang [HeTa26] raised the lower bound to $B>0.5850724$, and the site records a slight further improvement to $B>0.5850788$ found by GPT (prompted by Sothanaphan) — direct evidence that machine-optimized constructions move this record. Closely related to Erdős #227 (erdosproblems.com/227) on the ratio of maximal term to maximum modulus along all radii. The attacker's tool: design explicit coefficient sequences (gap-series-style constructions) and certify $\liminf \mu(r)/M(r)$ rigorously via interval arithmetic to push the lower bound past the current record, or sharpen the Clunie-type upper-bound argument into an explicit constant $c$.
References
| Ref | Source | Type |
|---|---|---|
| REF-01 | Erdős Problem #513 (T. F. Bloom) | website |
| REF-02 | Erdős Problem #227 — related maximal-term vs maximum-modulus problem | website |
Investigations · 0
No published investigations yet. This problem is unclaimed territory.