Wiki
Wiki

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

Updated


Statement

Let N>0N>0. If congruence classes with distinct moduli di>1d_i>1, each dividing NN, cover a full period of length NN, then

2N≤σ1(N),σ1(N)=∑d∣Nd.2N\le\sigma_1(N),\qquad \sigma_1(N)=\sum_{d\mid N}d.

In particular the least common multiple LL of the moduli of any finite covering of Z\mathbb Z with distinct moduli greater than one is non-deficient. If all the moduli are odd, LL is odd. Here non-deficient means perfect or abundant; the conclusion is not automatically strict abundance.

Complete proof

By Lemma 4.1, N≤∑iN/diN\le\sum_i N/d_i. Distinctness makes the did_i a subset of the divisors of NN other than 11. Hence

N≤∑iNdi≤∑d∣N, d>1Nd=σ1(N)−N.N\le\sum_i\frac N{d_i} \le\sum_{d\mid N,\ d>1}\frac Nd =\sigma_1(N)-N.

The last identity follows because d↦N/dd\mapsto N/d is an involution of the positive divisors, and the removed divisor 11 contributes NN. Rearranging proves the claim.

For an integer covering, L>0L>0 because it is the least common multiple of positive integers; every modulus divides it. The periodicity lemma reduces to its finite period. A covering has a nonempty index set, although the empty least common multiple convention L=1L=1 is harmless. If every modulus is odd, their product is odd and is divisible by LL. Thus LL is odd too.

Source and dependencies

Canonical v1, p. 5, Lemma 4.2 and §4.2. Complete deduction from Lemma 4.1 and elementary divisor arithmetic. The authors describe this density observation as folklore, without a novelty claim.

Bears on