For positive integers a matching (i.e. dividing them into pairs) is called {\it non-square} if the product of two numbers in each pair is not a perfect square. Prove that if there is a non-square matching, then there are at least non-square matchings.
(By denote the product .)
Problem 1631
Official solution
1. Define Good and Bad Pairs:
We define a pair as "good" if is a perfect square, and "bad" otherwise. Let our initial numbers be . Suppose we have a non-square matching .
2. Key Claim:
If we take any two pairs, say the first two pairs, then either both and are good, or both and are good.
3. Proof of the Claim:
Assume the contrary. Without loss of generality (WLOG), suppose is bad. If both and are bad, then must be good, which contradicts the assumption that is part of a non-square matching. Similarly, if both and are bad, then must be good, which again contradicts the assumption. Therefore, the claim is proved.
4. Induction Hypothesis:
We will use induction to prove the statement. The base case for is trivial since there is only one pair, and it is given to be a non-square matching.
5. Inductive Step:
Assume the statement is true for . From the claim, for each good pair , we can create new good pairs such that .
6. Constructing New Matchings:
Let's make a new match for from the set , . The remaining numbers can form at least matchings. Thus, we get new matchings, which equals .
7. Fixing Pairs and Continuing:
Now fix and do the same for . Similarly, we will get at least new matchings. Continue this process until the end. So we will get at least new matchings.
8. Adding the Initial Matching:
Including the initial matching , we get at least matchings.
The final answer is