Maths Olympiad Prep

Track / Stage 6 / 260 of 400 #1740 of 2444

Problem 1740

National Olympiad, first round
Combinatorics Difficulty 6.6 Prove it Olimpiade Italiana di Matematica · Italy

Let nn be an integer greater than 22. We want to color red exactly n+1n+1 of the numbers 1,2,,2n1,2n1, 2, \cdots, 2n-1, 2n in such a way that there are no three distinct numbers x,y,zx, y, z colored red satisfying the equality x+y=zx+y=z. Prove that there exists one and only one way to choose the numbers to color red that respects the given condition.

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

Solution:

Let AA be the set of numbers to color red. If A={n,n+1,,2n}A = \{n, n+1, \ldots, 2n\} there are no three of them for which x+y=zx+y=z, since for every x,yAx, y \in A with xyx \neq y we have x+y>2nx+y > 2n.

It remains therefore to prove that this is the only possible choice for the set AA.

We prove the statement by induction on nn, beginning with the case n=3n=3. If, for the sake of contradiction, we had 1A1 \in A then AA could not have two consecutive numbers greater than 11, so necessarily A={1,2,4,6}A = \{1, 2, 4, 6\}. But, since 2+4=62+4=6, this gives a contradiction. If instead 1A1 \notin A but 2A2 \in A, then AA would have to contain at least one of the two pairs {3,5}\{3, 5\} or {4,6}\{4, 6\}, again giving a contradiction with the hypothesis that one cannot have x+y=zx+y=z with x,y,zAx, y, z \in A.

Suppose now that we have proven uniqueness for the number nn and let us prove it for n+1n+1.

First of all we observe that AA must contain at least one of the numbers 2n+1,2n+22n+1, 2n+2, since otherwise AA would have to contain n+2n+2 numbers between 11 and 2n2n, which is excluded by the inductive hypothesis (only n+1n+1 numbers between 11 and 2n2n can belong to AA).

We also show that AA must contain both numbers 2n+12n+1 and 2n+22n+2: if this were not the case, AA would have to contain at least n+1n+1 of the numbers {1,,2n}\{1, \ldots, 2n\} and, by the inductive hypothesis, would have to contain {n,n+1,,2n}\{n, n+1, \ldots, 2n\}. But this would mean that AA contains neither n+(n+1)=2n+1n+(n+1)=2n+1 nor n+(n+2)=2n+2n+(n+2)=2n+2, a contradiction.

At this point, since 2n+1A2n+1 \in A, AA can contain only one of the numbers from the pairs {1,2n},{2,2n1},,{n,n+1}\{1, 2n\}, \{2, 2n-1\}, \ldots, \{n, n+1\}. One sees immediately that 1A1 \notin A since 1+(2n+1)=2n+21+(2n+1)=2n+2, and therefore 2nA2n \in A. Similarly, 2A2 \notin A, since 2+2nA2+2n \in A and therefore 2n1A2n-1 \in A. Inductively, for every (a,b)(a, b) with a+b=2n+1a+b=2n+1, the larger of the elements of the pair must necessarily belong to AA, and therefore A={n+1,n+2,,2n+2}A = \{n+1, n+2, \ldots, 2n+2\}.

Source: MathNet, licensed CC-BY-4.0. Statement translated into English from it; metadata (topic, difficulty, ordering) added by this project.