Wiki
Wiki

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

Updated


Statement

Setting. U\mathcal U is the set of positive odd integers not of the form p+2kp+2^k with pp prime and kk a positive integer (pp. 1--2); ω(m)\omega(m) is the number of distinct prime divisors of mm.

Theorem 1.3 (p. 2). min⁡m=11184810\min m=11184810 and min⁡ω(m)=7\min\omega(m)=7, and ω(m)=7\omega(m)=7 if and only if m=11184810m=11184810, where both minima are taken over all infinite arithmetic progressions {mh+a:h=0,1,…}\{mh+a:h=0,1,\ldots\} for which {mh+a:h=0,1,…}∖U\{mh+a:h=0,1,\ldots\}\setminus\mathcal U has asymptotic density zero.

Corollary 1.4 (p. 2). The same three conclusions hold with the minima taken over all infinite arithmetic progressions {mh+a:h=0,1,…}⊆U\{mh+a:h=0,1,\ldots\}\subseteq\mathcal U.

Here 11184810=2⋅3⋅5⋅7⋅13⋅17⋅24111184810=2\cdot3\cdot5\cdot7\cdot13\cdot17\cdot241. The paper notes (p. 3) that the value min⁡m=11184810\min m=11184810 in Corollary 1.4 was obtained earlier by Chen, Dai and Li (arXiv:2402.06644) through a heavy calculation.

Source. Yong-Gao Chen, A conjecture of Erdős on p+2kp+2^k, arXiv:2312.04120v3 (2024). Labels and pages are those of arXiv v3: the statements on p. 2, the proofs in Section 4 (Lemmas 4.1--4.4 on pp. 17--23, the proofs of Theorem 1.3 and Corollary 1.4 on p. 23). The edition read is identified on the source card.

Read depth. Claims checked: the statements were read clause by clause on the printed pages, and the factorization of 11184810 was checked. The proof was read but not checked step by step; the paper omits the proof of Lemma 4.2, saying it can be verified directly. Nothing here is independently reviewed.

Proof pointer

Section 4. If {mh+a}∖U\{mh+a\}\setminus\mathcal U has density zero then mm is even and aa odd, and by a positive-proportion result of Sun (Lemma 4.4, p. 23) (a−2ℓ,m)>1(a-2^\ell,m)>1 for every ℓ≥1\ell\ge1. The odd primes p∣mp\mid m with a≡2ℓ(modp)a\equiv2^\ell\pmod p for some ℓ\ell then form a "well constructed prime set": the residues ℓ≡ai(modrpi)\ell\equiv a_i\pmod{r_{p_i}}, with rpr_p the order of 2 modulo pp, cover the integers. Lemma 4.3 (p. 18, proof to p. 22), a case analysis on minimal covering systems using the necessary condition ∑1/mi≥1\sum1/m_i\ge1, Lemma 4.1 and the table of small orders in Lemma 4.2, shows such a set has at least six primes with product at least 55924055592405, and exactly six if and only if the product is 55924055592405. Attainment comes from Lemma 3.3 (p. 14): {11184810s+992077:s≥0}⊆U\{11184810s+992077:s\ge0\}\subseteq\mathcal U.

Dependencies

Lemma 4.4, quoted from Sun; Lemmas 4.1--4.3 on covering systems and orders of 2; Lemma 3.3 for the extremal progression.

Bears on

  • Problem 16: the paper derives from Theorems 1.3 and 1.5 a third proof (p. 25) that the answer to the problem is no; see Theorem 3.1.