Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Source. Theorem 1.1, p. 2, proved in Section 3 on pp. 3--6, of Carl Pomerance, On amicable numbers, Analytic Number Theory: In Honor of Helmut Maier's 60th Birthday, Springer (2015), 321--327, doi:10.1007/978-3-319-22240-0_19, as identified on the source card. Labels and pages are those of the author's manuscript named there (pp. 1--7).
Statement
Notation (pp. 1--2). is the sum-of-divisors function and . Two different positive integers with and form an amicable pair; a positive integer is amicable if it belongs to an amicable pair. is the set of amicable numbers and .
Theorem 1.1 (p. 2, quoted). "As , we have"
Since , this gives for all sufficiently large , the form stated in the abstract (p. 1). The paper presents the theorem as replacing the exponent in the earlier bound of Pomerance's 1981 paper (its reference [15]) by (p. 2).
Proof pointer
Section 3, pp. 3--6. With , the proof discards, in steps (i)--(vii), sets of of size at most : those with or small, with a large -smooth divisor or a large squarefull divisor, with a prime above dividing , with small cofactors of the largest primes of and , with or above (the new case, treated through a congruence modulo for a suitable divisor of ), and with for the largest squarefree unitary divisor of (or likewise of ), which Lemma 2.1 controls. For the remaining , a prime dividing must also divide for a prime power dividing , and summing over gives (p. 6).
Dependencies
Lemma 2.1; de Bruijn's bound for smooth numbers (reference [3]); the elementary inequality (1), p. 4, taken from the proofs of [11, Lemma 3.6] and [12, Lemma 3.3]. Read depth: claims checked; the statement was read clause by clause on p. 2, the proof for its structure on pp. 3--6.
Bears on
- #830: each amicable pair is determined by its smaller member , so the theorem bounds from above the number of pairs with that the problem's counts. The problem's condition also admits , that is perfect numbers, which the paper does not count. The bound is of the form , so it is consistent with the conjectured ; the paper answers neither question of the problem and notes that the infinitude of amicable pairs is unproved (p. 1).