Olympiad Maths Prep

Track / Stage 7 / 21 of 300 #1421 of 2000

Problem 1421

National olympiad second round; IMO P1/P4
Number theory Difficulty 7.0 Prove it

The sequence of integers (an)n1(a_{n})_{n \ge 1} is defined by a1=2a_{1}= 2, an+1=2an21a_{n+1}= 2a_{n}^{2}-1.
Prove that for each positive integer n, nn and ana_{n} are coprime.

This one wants a proof. Work it on paper, read the official solution, then mark yourself honestly — the ladder only means something if the record is true.

Official solution

To prove that for each positive integer n n , n n and an a_n are coprime, we will use induction and properties of the sequence defined by a1=2 a_1 = 2 and an+1=2an21 a_{n+1} = 2a_n^2 - 1 .

1. Base Case:
For n=1 n = 1 , a1=2 a_1 = 2 . Clearly, gcd(1,2)=1 \gcd(1, 2) = 1 , so 1 1 and a1 a_1 are coprime.

2. Inductive Step:
Assume that for some k1 k \geq 1 , k k and ak a_k are coprime, i.e., gcd(k,ak)=1 \gcd(k, a_k) = 1 . We need to show that k+1 k+1 and ak+1 a_{k+1} are also coprime.

Given ak+1=2ak21 a_{k+1} = 2a_k^2 - 1 , we need to show that gcd(k+1,2ak21)=1 \gcd(k+1, 2a_k^2 - 1) = 1 .

3. Using the Inductive Hypothesis:
Suppose p p is a prime divisor of both k+1 k+1 and ak+1 a_{k+1} . Then p2ak21 p \mid 2a_k^2 - 1 . Since pk+1 p \mid k+1 , we have k+10(modp) k+1 \equiv 0 \pmod{p} .

4. Residue Analysis:
Let bi=aimodp b_i = a_i \mod p be the residue of ai a_i modulo p p . We need to analyze the sequence bi b_i modulo p p .

- If bk=0 b_k = 0 , then ak0(modp) a_k \equiv 0 \pmod{p} . This implies ak+1=2ak211(modp) a_{k+1} = 2a_k^2 - 1 \equiv -1 \pmod{p} , so bk+11(modp) b_{k+1} \equiv -1 \pmod{p} . Hence, pak+1 p \nmid a_{k+1} .
- If bk=1 b_k = 1 , then ak1(modp) a_k \equiv 1 \pmod{p} . This implies ak+1=2ak211(modp) a_{k+1} = 2a_k^2 - 1 \equiv 1 \pmod{p} . Hence, pak+1 p \nmid a_{k+1} .

5. Pigeonhole Principle:
If bk0,1 b_k \neq 0, 1 for 1kp1 1 \leq k \leq p-1 , by the pigeonhole principle, there must exist indices 1i<jp1 1 \leq i < j \leq p-1 such that bi=bj b_i = b_j . This implies that the sequence bi b_i is periodic with period ji j - i without ever being zero. Hence, pak+1 p \nmid a_{k+1} .

6. Conclusion:
Since pak+1 p \nmid a_{k+1} for any prime p p dividing k+1 k+1 , it follows that gcd(k+1,ak+1)=1 \gcd(k+1, a_{k+1}) = 1 .

By induction, we have shown that for all n1 n \geq 1 , gcd(n,an)=1 \gcd(n, a_n) = 1 .

\blacksquare

Source: NuminaMath-1.5, licensed Apache-2.0. Statement and solution reproduced as published; topic, difficulty and ordering added by this site.