Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Ahlswede 1996 sets integers pairwise common divisor factor
Ahlswede, Rudolf and Khachatrian, Levon H., Sets of integers with pairwise common divisor and a factor from a specified set of primes. Acta Arith. 75 (1996), 259--276.
For a finite prime set Q = {q_1 < ... < q_r}, let I(n,Q) consist of the sets A of integers up to n such that any two elements have a common divisor greater than 1 and every element shares a factor with the product of Q, and let f(n,Q) be the largest cardinality. Theorem 1, the main result, evaluates it: for n at least the product of the q_i, f(n,Q) = max over 1 <= j <= r of |M(2q_1, ..., 2q_j, q_1...q_j) cap N(n)|, where M(.) is the set of multiples. Specializing to n = q_1^{a_1} ... q_r^{a_r} recovers the Erdos-Graham problem on the maximal size g(n) of a set 1 < a_1 < ... < a_k = n with all pairs non-coprime, so Theorem 1 in particular proves the conjecture Erdos formulated (Conjecture 1) after the authors told him, during his 1992 visit to Bielefeld, that the earlier published guess for g(n) has easy counterexamples. A corollary computes the maximal upper asymptotic density of an infinite such set as max over j of (1/2)(1 - prod_{i<=j}(1 - 1/q_i) + 1/(q_1...q_j)), attained by a set possessing an asymptotic density. Theorem 2 gives the same formula among squarefree integers, f*(n,Q) = max over j of |M(2q_1, ..., 2q_j, q_1...q_j) cap N*(n)|, with no restriction on n; Section 5 only sketches its proof. Section 6 shows the lower bound restriction on n in Theorem 1 cannot be dropped. The setting is dual to that of their paper Sets of integers and quasi-integers with pairwise common divisor (Acta Arith. 74 (1996), 141--153), where a prime set is excluded rather than required. For Erdos problem 534 this settles the extremal value for sets with pairwise common divisors and a factor from a prescribed prime set whenever n is at least the product of the primes, which holds in the specialization above, so the problem's maximum is the value Conjecture 1 predicts.
Source: http://www.impan.pl/get/doi/10.4064/aa-75-3-259-276. The file's text layer carries no copyright or license line, and the publisher's record (https://www.impan.pl/get/doi/10.4064/aa-75-3-259-276, read 2026-10-02) offers the PDF under the link "Pobierz zgodnie z CC-BY" ("Free download under CC-BY license" on the English site), a Creative Commons Attribution license whose version the record does not name.
Bears on. #534
Results to transcribe.
- Theorem 1: For every finite Q = {q_1 < ... < q_r} of primes and n >= prod q_i, f(n,Q) = max_{1<=j<=r} |M(2q_1, ..., 2q_j, q_1...q_j) cap N(n)|; in particular Erdos's Conjecture 1 for g(n) is true.
- Corollary: The maximal upper asymptotic density of an infinite admissible set is max_{1<=j<=r} (1/2)(1 - prod_{i=1}^{j}(1 - 1/q_i) + 1/(q_1...q_j)), and the maximum is attained by a set with an asymptotic density.
- Theorem 2: For every finite Q = {q_1 < ... < q_r} of primes and every n, f*(n,Q) = max_{1<=j<=r} |M(2q_1, ..., 2q_j, q_1...q_j) cap N*(n)|, where f*(n,Q) is the largest size of such a set of squarefree integers up to n; the proof is sketched in Section 5.
- Section 6 remark: The restriction n >= prod_{i} q_i in Theorem 1 cannot be ignored.