Maths Olympiad Prep

Track / Stage 8 / 28 of 180 #2368 of 2604

Problem 2368

IMO Shortlist mid-range; USAMO P2/P5
Number theory Difficulty 8.1 Prove it Team Selection Test for IMO · North Macedonia

Let aa and nn be integers. We define an=1+a+a2+⋯+an−1a_n = 1 + a + a^2 + \dots + a^{n-1}. Prove that if ap≡1(modp)a^p \equiv 1 \pmod p for every prime divisor pp of n2−n1n_2 - n_1, then the number an2−an1n2−n1\frac{a_{n_2} - a_{n_1}}{n_2 - n_1} is an integer.

This one wants a proof. Work it on paper, then read the official solution and mark yourself. Be honest about it: the record is only any use to you if it is.

Next problem →

Official solution

Lemma. Let aa and nn be integers such that a≡1(modp)a \equiv 1 \pmod p for each prime p∤np \nmid n and an=1+a+a2+⋯+an−1a_n = 1 + a + a^2 + \dots + a^{n-1}. Then n∣ann \mid a_n.

Proof of the lemma. Let prp^r be the largest power of the prime number pp such that pr∣np^r \mid n. We will prove the equality
1+a+a2+⋯+an−1=(1+apr+a2pr+⋯+a(p−1)pr)∏k=1r(1+apk−1+a2pk−1+⋯+a(p−1)pk−1) 1+a+a^2+\dots+a^{n-1} = \left(1+a^{p^r}+a^{2p^r}+\dots+a^{(p-1)p^r}\right) \prod_{k=1}^{r} \left(1+a^{p^{k-1}}+a^{2p^{k-1}}+\dots+a^{(p-1)p^{k-1}}\right)
for each integer aa. If a=1a=1 the left-hand side is nn and the right-hand side is nprpr=n\frac{n}{p^r}p^r = n (one pp for each term in the product). Let a≠1a \neq 1. If we multiply the left-hand side and right-hand side by a−1a-1 from the right we get
(a−1)(1+a+⋯+ap−1)(1+ap+a2p+⋯+a(p−1)p)…(1+a2pr+⋯+apr)=(ap−1)(1+ap+a2p+⋯+a(p−1)p)…(1+a2pr+⋯+apr)=(ap2−1)(1+ap2+⋯+a(p−1)p2)…(1+apk−1+⋯+apk−1)=(apr−1)(1+apr+⋯+apr−1)=an−1 \begin{aligned} & (a-1)(1+a+\dots+a^{p-1})(1+a^p+a^{2p}+\dots+a^{(p-1)p})\dots \left(1+a^{2p^r}+\dots+a^{p^r}\right) \\ &= (a^p-1)(1+a^p+a^{2p}+\dots+a^{(p-1)p})\dots \left(1+a^{2p^r}+\dots+a^{p^r}\right) \\ &= (a^{p^2}-1)(1+a^{p^2}+\dots+a^{(p-1)p^2})\dots \left(1+a^{p^{k-1}}+\dots+a^{p^{k-1}}\right) \\ &= (a^{p^r}-1)\left(1+a^{p^r}+\dots+a^{p^{r-1}}\right) = a^n-1 \end{aligned}
Each expression in the product is divisible by pp since
1+apk−1+a2pk−1+⋯+a(p−1)pk−1=(apk−1−1)+(a2pk−1−1)+⋯+(a(p−1)pk−1−1)+p 1+a^{p^{k-1}}+a^{2p^{k-1}}+\dots+a^{(p-1)p^{k-1}} = (a^{p^{k-1}}-1)+(a^{2p^{k-1}}-1)+\dots+(a^{(p-1)p^{k-1}}-1)+p
each of the expressions in brackets is divisible by pp.

Without loss of generality we can assume that n1<n2n_1 < n_2. It is clear that
an2−an1n2−n1=1+a+⋯+an2−1−1−a−⋯−an1−1n2−n1=an1(1+a+⋯+an2−n1−1)n2−n1 \frac{a_{n_2} - a_{n_1}}{n_2 - n_1} = \frac{1 + a + \dots + a^{n_2-1} - 1 - a - \dots - a^{n_1-1}}{n_2 - n_1} = \frac{a^{n_1}(1 + a + \dots + a^{n_2-n_1-1})}{n_2 - n_1}
Using the lemma we get (n2−n1)∣an2−n1(n_2 - n_1) \mid a_{n_2 - n_1} from where we get that the number an2−an1n2−n1\frac{a_{n_2} - a_{n_1}}{n_2 - n_1} is a natural number.

Source: MathNet, licensed CC-BY-4.0. Statement reproduced verbatim; metadata (topic, difficulty, ordering) added by this project.