Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Source: arXiv v2, p. 3, Theorem 3; proof on pp. 6--7.
Statement
There is an absolute constant such that every finite covering system of multiplicity has smallest modulus at most
Full proof relative to the stated external inputs
If the family contains modulus , (1) is immediate. Otherwise write its moduli as and use the notation of the distortion setup.
First consider large . Put , where the absolute constant will be fixed below. Let be the largest index with , with the convention if no such prime divides . Set
Using the first term in the minimum when and noting that the second denominator is when , define
By Lemma 3.3(b), followed by the standard Chebyshev upper bound and partial summation,
The supremum of the last expression over tends to zero as . Fix so large that .
Choose a preliminary absolute constant and set
and suppose for a contradiction that . Lemma 3.3(a) gives
For all sufficiently large , one has and . Moreover
Consequently . Choose sufficiently large. The external smooth-number estimate in Lemma 3.4 then makes the sum in (4) less than for all sufficiently large . Thus and .
Choose so that the preceding argument applies whenever . For completeness, the finitely many multiplicities can be handled by the same criterion without an asymptotic assertion. Choose large, set
and take the cutoff . For large , and
because . The right side of (3), with and , tends to zero. At the same time, Lemma 3.4 bounds (4), with , by
which also tends to zero. Hence one common finite bounds the smallest modulus for every . Since in this finite range, choose the final constant so that the right side of (1) is at least for all these . For , an assumed violation of (1) also gives , so the preceding large- argument with the unchanged threshold still applies.
We have therefore arranged, for every and under an assumed violation of (1),
The exact external distortion criterion says that then fails to cover , a contradiction. Thus (1) holds.
The printed proof gives the asymptotic large- calculation and leaves the finite range implicit; the common- paragraph supplies that routine closure. Also, its final asymptotic for omits the factor coming from . The corrected relation (5) has the same consequence after enlarging the unspecified absolute constant .
Bears on
- Problem 2. At the theorem bounds the least modulus of every covering system with distinct moduli by an unspecified absolute constant; it is not a new explicit improvement on the density paper's bound.