OpenFrontiers
← Back to the index
OP-12051OpenNumber theory

Asymptotic Dobrowolski constant for Lehmer’s problem (C40bC_{40b})

Asymptotic Dobrowolski constant for Lehmer’s problem: determine C40bC_{40b} or improve its bounds; the repository records 9/49/4 (lower) and ∞\infty (upper).

Editorial MSC2020 classification. Original article codes and source evidence are recorded in the literature directory.

Licensed source statementAs of Repository table at commit 2c1968c (2026-09-26); open while the recorded lower and upper bounds differSource & evidence ↓

Problem statement

Let α\alpha be a nonzero algebraic number of degree dd, with minimal polynomial over Z\mathbb{Z}

f(X)=ad∏i=1d(X−αi),f(X)=a_d\prod_{i=1}^d (X-\alpha_i),

where ad>0a_d>0 and α1,…,αd\alpha_1,\dots,\alpha_d are the conjugates of α\alpha. Define the Mahler measure of α\alpha by

M(α)  :=  ad∏i=1dmax⁡{1,∣αi∣}.M(\alpha)\;:=\;a_d\prod_{i=1}^d \max\{1,\lvert \alpha_i\rvert\}.

[Vou1996-def-M]

Define the absolute logarithmic height h(α)h(\alpha) by

h(α)  :=  log⁡M(α)d.h(\alpha)\;:=\;\frac{\log M(\alpha)}{d}.

[Vou1996-def-h]

Write

B(d)  :=  (log⁡log⁡dlog⁡d)3(d≥3),B(d)\;:=\;\left(\frac{\log\log d}{\log d}\right)^3 \qquad (d\ge 3),

and consider algebraic numbers α\alpha that are not roots of unity. Dobrowolski proved an asymptotic lower bound of the form

M(α)  >  1+(1−ϵ) B(d)for d≥d(ϵ),M(\alpha)\;>\;1+(1-\epsilon)\,B(d)\quad \text{for }d\ge d(\epsilon),

for each ϵ>0\epsilon>0. [Vou1996-dob-asymp]

Motivated by this asymptotic form, define the asymptotic Dobrowolski constant C40bC_{40b} to be the largest constant cc such that, for every ϵ>0\epsilon>0, there exists d(ϵ)d(\epsilon) with

M(α)  ≥  1+(c−ϵ) B(d)for all non-root-of-unity α of degree d≥d(ϵ).M(\alpha)\;\ge\;1+(c-\epsilon)\,B(d)\quad \text{for all non-root-of-unity }\alpha\text{ of degree }d\ge d(\epsilon).

Lean 4

Not yet attached

No Lean statement or proof is attached yet. Have a formalization? Share its source to add it to this problem.

Background

  • M(α)M(\alpha) vs. log⁡M(α)\log M(\alpha). Many statements in the literature are formulated for log⁡M(α)\log M(\alpha) (equivalently d h(α)d\,h(\alpha)). Voutier’s inequality
1+log⁡(M(α))<M(α)1+\log(M(\alpha))<M(\alpha)

shows that any bound of the form log⁡M(α)>c B(d)\log M(\alpha) > c\,B(d) immediately implies the corresponding bound M(α)>1+c B(d)M(\alpha) > 1+c\,B(d). [Vou1996-log-vs-M]

  • Earlier degree-dependent lower bounds of different shape. Before Dobrowolski’s (log⁡log⁡d/log⁡d)3(\log\log d/\log d)^3 term, Blanksby–Montgomery proved
M(α)  >  1+152dlog⁡(6d),M(\alpha)\;>\;1+\frac{1}{52d\log(6d)},

and Stewart proved

M(α)  >  1+1104dlog⁡(d).M(\alpha)\;>\;1+\frac{1}{10^4d\log(d)}.

[Vou1996-bm] [Vou1996-stew]

References (from the repository page)

  • [Vou1996] Voutier, Paul M. An effective lower bound for the height of algebraic numbers. Acta Arithmetica 74(1) (1996), 81–95. DOI: 10.4064/aa-74-1-81-95. Google Scholar. arXiv PDF.

  • [Vou1996-def-M] loc: arXiv v1 PDF p.1, Introduction (definition of M(α)M(\alpha)). quote: “We shall define the Mahler measure of α\alpha, M(α)M(\alpha), by M(α)=ad∏i=1dmax⁡(1,∣αi∣)M(\alpha)=a_d\prod_{i=1}^d\max(1,\lvert\alpha_i\rvert).”

  • [Vou1996-def-h] loc: arXiv v1 PDF p.1, Introduction (definition of h(α)h(\alpha)). quote: “h(α)=log⁡M(α)dh(\alpha)=\frac{\log M(\alpha)}{d}.”

  • [Vou1996-bm] loc: arXiv v1 PDF p.1, Introduction (Blanksby–Montgomery). quote: “They proved that M(α)>1+152dlog⁡(6d)M(\alpha)>1+\frac{1}{52d\log(6d)}.”

  • [Vou1996-stew] loc: arXiv v1 PDF p.1, Introduction (Stewart). quote: “In 1978, C.L. Stewart [18] introduced a method from transcendental number theory to prove that M(α)>1+1/(104dlog⁡(d))M(\alpha) > 1+1/(10^4d \log(d)).”

  • [Vou1996-dob-asymp] loc: arXiv v1 PDF p.2, Introduction (Dobrowolski’s asymptotic bound). quote: “Dobrowolski… showed that M(α)>1+(1−ϵ)(log⁡log⁡dlog⁡d)3M(\alpha)>1+(1-\epsilon)\left(\frac{\log\log d}{\log d}\right)^3 for d≥d(ϵ)d\ge d(\epsilon).”

  • [Vou1996-cs-lou] loc: arXiv v1 PDF p.2, Introduction (Cantor–Straus; Louboutin). quote: “Cantor and Straus… replace the coefficient (1−ϵ)(1-\epsilon) by (2−ϵ)(2-\epsilon). Louboutin… to (94−ϵ)(\frac94-\epsilon).”

  • [Vou1996-log-vs-M] loc: arXiv v1 PDF p.3, paragraph after Theorem. quote: “Notice that 1+log⁡(M(α))<M(α)≤∣α∣d1+\log(M(\alpha))<M(\alpha)\le\lvert\alpha\rvert^d.”

  • [Dob1979] Dobrowolski, E. On a question of Lehmer and the number of irreducible factors of a polynomial. Acta Arithmetica 34 (1979), 391–401. Google Scholar.

  • [CS1982] Cantor, D.; Straus, E. G. On a conjecture of D. H. Lehmer. Acta Arithmetica 42(1) (1982), 97–100. Google Scholar.

  • [Lou1983] Louboutin, R. Sur la mesure de Mahler d'un nombre algebrique. C. R. Acad. Sci. Paris Ser. I 296 (1983), 707–708. Google Scholar.

Progress & source assessment

Repository table (commit 2c1968c, 26 September 2026): best known lower bound 9/49/4; best known upper bound ∞\infty. Values marked with an asterisk have minimal verification.

Known upper bounds

Bound Reference Comments
∞\infty

Known lower bounds

Bound Reference Comments
11 [Dob1979] Dobrowolski proved M(α)>1+(1−ϵ)B(d)M(\alpha) > 1+(1-\epsilon)B(d) for d≥d(ϵ)d\ge d(\epsilon) (as reported by Voutier), hence C40b≥1C_{40b}\ge 1. [Vou1996-dob-asymp]
22 [CS1982] Cantor–Straus replace the coefficient (1−ϵ)(1-\epsilon) by (2−ϵ)(2-\epsilon) (as reported by Voutier), hence C40b≥2C_{40b}\ge 2. [Vou1996-cs-lou]
9/49/4 [Lou1983] Louboutin improves the coefficient to (94−ϵ)(\tfrac94-\epsilon) (as reported by Voutier), hence C40b≥9/4C_{40b}\ge 9/4. [Vou1996-cs-lou]
References
  1. [1]Davis, Damek; Ivanisvili, Paata; Tao, Terence (maintainers). Optimization Constants in Mathematics, github.com/teorth/optimizationproblems (commit 2c1968c, 26 September 2026). — Asymptotic Dobrowolski constant for Lehmer’s problem ($C_{40b}$) linkApache-2.0 (repository license); constant pages reproduced with attribution; repository references retained.

Recorded progress

Add yours →

Nothing recorded yet. A pointer to a paper the entry is missing counts.

Machine attempts

0 recorded

Attempts run against the statement below and are recorded in full, including the failures. The point is to stop the next model — and the next person — from re-walking a road that is already known to end. Each attempt costs 25 credits.

No machine attempts recorded.

Sign in to run an attempt. New accounts start with enough credits for ten.

Discussion

0 notes

Notes are for clarifying the statement, flagging a missing reference, or recording why an approach fails. Claimed solutions belong in a submission, where they can be reviewed.

No notes yet.

Sign in to add a note.