Problem 1 from the IMO 2009
Here is Problem 1 from the IMO 2009:
| |
Let n be a positive integer and let a1, ..., ak (k ≥ 2) be distinct integers in the set {1, ..., n} such that n divides ai(ai+1 - 1) for i = 1, ..., k - 1. Prove that n does not divide ak(a1 - 1).
|
Solution
|Activities|
|Contact|
|Front page|
|Contents|
|Algebra|
|Store|
Copyright © 1996-2012 Alexander Bogomolny
Set for simplicity a = a1 and b = a2, a ≠ b and both are less than n.
Thus n|a(b - 1) meaning that there are s and t, st = n, such that s|a and t|(b - 1). Since both a and b are less than n, neither s or t may equal 1, meaning that both are non-trivial factors of n.
Let first k = 2, and assume that n|b(a - 1). No non-trivial factor of t may divide b, and no non-trivial factor of s may divide a - 1, implying that s|b. We are therefore in a position to apply the Chinese Remainder Theorem:
| (1) |
| b | = 1 (mod t), |
| b | = 0 (mod s). |
|
The theorem tells us that there is a unique solution modulo lcd(s, t) = n. However, due to the symmetry in the conditions for a and b, we also have
| | a | = 1 (mod t), |
| | a | = 0 (mod s), |
which contradicts the assumption that a and b are different.
For k = 3, we are given that n|b(c - 1) where c = a3. Assume, as before, n|c(a - 1).
There are s and t different from 1 and n such that s|a and t|(b - 1). No non-trivial factor of t may divide b. From n|b(c - 1) it then follows that t|(c - 1). If not s itself, then one of its factors, say, s0 divides b: s0|b, s1t|(c - 1), where s = s0s1.
Similarly, n|c(a - 1) implies that σ0|c and σ1s1t|(a - 1), where σ0σ1 = s0. It's now one of the two, either σ1s1 = 1 or not. In the former case, the Chinese Remainder theorem leads to a contradiction. In the latter case, there appears a non-trivial factor common to both a and and a - 1. Again a contradiction.
For greater k, we always have t|ai+1 - 1. As we progress from 1 to larger i, additional factors may be borrowed from s. If this happens, then a and a - 1 will have shown to have common factors. Otherwise a contradiction is reached via the Chinese Remainder theorem.
A shorter and a more formal solution was posted at the artofproblemsolving.com forum by mathman145.
We know that n divides ai(ai + 1 - 1), i = 1, ..., k - 1. The condition is equivalent to saying that
| | akak - 1 = ak - 1 (mod n), |
| | akak - 1ak - 2 = ak - 1ak - 2 = ak - 2 (mod n), |
| | akak - 1ak - 2ak - 3 = ak - 1ak - 2ak - 3 = ak - 2ak - 3 = ak - 3 (mod n) ... |
so that
| | akak - 1ak - 2ak - 3 ... a1 = a1 (mod n). |
Now assume that ak(a1 - 1) is divisible by n. Then, similarly to the above
| | a1akak - 1ak - 2ak - 3 ... a2 = a2 (mod n), |
implying that a1 = a2 (mod n) which is impossible for two distinct positive integers both less than n.
|Activities|
|Contact|
|Front page|
|Contents|
|Algebra|
|Store|
Copyright © 1996-2012 Alexander Bogomolny
|