Skip to main content
OlympiadHQ

Browse · MathNet

Print

2022 China Team Selection Test for IMO

China 2022 number theory

Problem

Given a positive integer , let be the set of positive divisors of , and let be a function. Prove that the following are equivalent: (A) for any positive divisor of , (B) for any positive divisor of ,
Solution
Proof: For map , we define a map to be By Möbius transform, the map is uniquely determined by : Therefore we have We need the following Lemma: If , then Assuming this lemma temporarily, we prove that (A) and (B) are equivalent. (The proof of the lemma will be given later.) “(B) ⇒ (A)” Suppose that for every we have . The lemma implies that

"(A) ⇒ (B)" Suppose that (A) holds. We use induction to prove that for , we always have . Suppose that we have already proved for and . Consider case of . By Lemma and the inductive hypothesis, we have From this we get , completing the induction.

It remains to prove the lemma. Note that Let be the prime factorization. Then we only need to prove for every . For this, we may assume that . Denote . Note that Then it suffices to show: if are multiples of , then Put . Then

Since is divisible by , we get So (2) holds, and we complete the proof of the lemma. Another proof of the lemma: We will use the following fact. Let be a finite set, and let be a map such that . Under the action of , is divided into some orbits. Each orbit looks like , where is the smallest positive integer such that ; we call the length of the orbit. Since , the length of every orbit is a divisor of . For every divisor of , let be the number of orbits with length . For every positive integer , let then . Therefore we get In particular, we have We apply this to the following model. Let be the set of all subsets of of elements. Define to be In this model, for every divisor of , we have Then (3) implies that

Techniques

Möbius inversionModular ArithmeticEnumeration with symmetryAlgebraic properties of binomial coefficients