Maths Olympiad Prep

Library / /115 of 383

Combinatorics Difficulty 8.2 Shortlist Prove it IMO

Let a1,a2,,a2023a_{1}, a_{2}, \ldots, a_{2023} be positive integers such that
- a1,a2,,a2023a_{1}, a_{2}, \ldots, a_{2023} is a permutation of 1,2,,20231,2, \ldots, 2023, and
- a1a2,a2a3,,a2022a2023\left|a_{1}-a_{2}\right|,\left|a_{2}-a_{3}\right|, \ldots,\left|a_{2022}-a_{2023}\right| is a permutation of 1,2,,20221,2, \ldots, 2022.
Prove that max(a1,a2023)507\max \left(a_{1}, a_{2023}\right) \geqslant 507.

Solution

For the sake of clarity, we consider and prove the following generalisation of the original problem (which is the case N=1012N=1012):
Let NN be a positive integer and a1,a2,,a2N1a_{1}, a_{2}, \ldots, a_{2N-1} be positive integers such that
- a1,a2,,a2N1a_{1}, a_{2}, \ldots, a_{2N-1} is a permutation of 1,2,,2N11,2, \ldots, 2N-1, and
- a1a2,a2a3,,a2N2a2N1\left|a_{1}-a_{2}\right|,\left|a_{2}-a_{3}\right|, \ldots,\left|a_{2N-2}-a_{2N-1}\right| is a permutation of 1,2,,2N21,2, \ldots, 2N-2.
Then a1+a2N1N+1a_{1}+a_{2N-1} \geqslant N+1 and hence max(a1,a2N1)N+12\max \left(a_{1}, a_{2N-1}\right) \geqslant\left\lceil\frac{N+1}{2}\right\rceil.

Now we proceed to the proof of the generalised statement. We introduce the notion of score of a number a{1,2,,2N1}a \in\{1,2, \ldots, 2N-1\}. The score of aa is defined to be
s(a):=aN. s(a):=|a-N| .
Note that, by the triangle inequality,
abaN+Nb=s(a)+s(b) |a-b| \leqslant|a-N|+|N-b|=s(a)+s(b)
Considering the sum a1a2+a2a3++a2N2a2N1\left|a_{1}-a_{2}\right|+\left|a_{2}-a_{3}\right|+\cdots+\left|a_{2N-2}-a_{2N-1}\right|, we find that
(N1)(2N1)=a1a2+a2a3++a2N2a2N12(s(a1)+s(a2)++s(a2N1))(s(a1)+s(a2N1))=2N(N1)(s(a1)+s(a2N1)) \begin{aligned} (N-1)(2N-1) & =\left|a_{1}-a_{2}\right|+\left|a_{2}-a_{3}\right|+\cdots+\left|a_{2N-2}-a_{2N-1}\right| \\ & \leqslant 2\left(s\left(a_{1}\right)+s\left(a_{2}\right)+\cdots+s\left(a_{2N-1}\right)\right)-\left(s\left(a_{1}\right)+s\left(a_{2N-1}\right)\right) \\ & =2N(N-1)-\left(s\left(a_{1}\right)+s\left(a_{2N-1}\right)\right) \end{aligned}
For the last equality we used that the numbers s(a1),s(a2),,s(a2N1)s\left(a_{1}\right), s\left(a_{2}\right), \ldots, s\left(a_{2N-1}\right) are a permutation of 0,1,1,2,2,,N1,N10,1,1,2,2, \ldots, N-1, N-1.
Hence, s(a1)+s(a2N1)2N(N1)(N1)(2N1)=N1s\left(a_{1}\right)+s\left(a_{2N-1}\right) \leqslant 2N(N-1)-(N-1)(2N-1)=N-1. We conclude that
(Na1)+(Na2N1)s(a1)+s(a2N1)N1, \left(N-a_{1}\right)+\left(N-a_{2N-1}\right) \leqslant s\left(a_{1}\right)+s\left(a_{2N-1}\right) \leqslant N-1,
which implies a1+a2N1N+1a_{1}+a_{2N-1} \geqslant N+1.

Want a route through all this instead of an archive? The track puts 2,000 problems in a working order, from AMC 10 level to the IMO shortlist.

Source: MathNet, licensed CC-BY-4.0. Statement reproduced verbatim; metadata (topic, difficulty) added by this project.