Maths Olympiad Prep

Library / /1 of 2

Algebra Difficulty 7.5 National Olympiad, round 2 Prove it Middle European Mathematical Olympiad (MEMO)

Problem:

Let kk be a positive integer and a1,a2,,aka_{1}, a_{2}, \ldots, a_{k} be nonnegative real numbers. Initially, there is a sequence of nkn \geq k zeros written on a blackboard. At each step, Nicole chooses kk consecutive numbers written on the blackboard and increases the first number by a1a_{1}, the second one by a2a_{2}, and so on, until she increases the kk-th one by aka_{k}. After a positive number of steps, Nicole managed to make all the numbers on the blackboard equal. Prove that all the nonzero numbers among a1,a2,,aka_{1}, a_{2}, \ldots, a_{k} are equal.

Solution

Solution:

Denote by Li,0i<nL_{i}, 0 \leq i < n, the number of tiles that John puts in such a way that a0a_{0} is added at position ii. Note that Lnk=Lnk+1==Ln1=0L_{n-k} = L_{n-k+1} = \cdots = L_{n-1} = 0. Analogously, we define Ri,0i<nR_{i}, 0 \leq i < n for the number of times ak1a_{k-1} was added at position ii.

First, note that a0L0=Ka_{0} L_{0} = K, since only the leftmost operations add something to the first zero. Note that for all 0i<k0 \leq i < k we have KaiL0=aiK/a0K \geq a_{i} L_{0} = a_{i} K / a_{0}, hence aia0a_{i} \leq a_{0}. Analogously, focusing on ak1a_{k-1} and Rn1R_{n-1} we conclude that for every 0i<k0 \leq i < k we have aiak1a_{i} \leq a_{k-1}. That is, we have a0=ak1a_{0} = a_{k-1} and this is the largest value in the sequence a0,,ak1a_{0}, \ldots, a_{k-1}; moreover, L0=Rn1L_{0} = R_{n-1}. To simplify notation, we denote a=a0=ak1a = a_{0} = a_{k-1} and p=L0=Rn1p = L_{0} = R_{n-1}.

Our task is to prove that every aia_{i} has value either 00 or aa. We will prove that by induction, in fact proving the following, stronger, statement. For every 0i<k/20 \leq i < \lceil k / 2 \rceil, we prove:

(a) ai=ak1ia_{i} = a_{k-1-i},

(b) Li=Rn1iL_{i} = R_{n-1-i},

(c) either ai=0a_{i} = 0 or ai=aa_{i} = a.

In plain words, besides proving the required statement, we are also gradually proving that the sequence aia_{i} is symmetrical, as well as John's moves.

We have already proven this statement for i=0i = 0. Let us now prove it for any ii, assuming it was proven for 0,,i10, \ldots, i-1.

Note that for the ii-th position on the blackboard we have
K=j=0iLjaij K = \sum_{j=0}^{i} L_{j} a_{i-j}
Let us first focus on the terms L1ai1,L2ai2,,Li1a1L_{1} a_{i-1}, L_{2} a_{i-2}, \ldots, L_{i-1} a_{1}. By the induction statement, the value of each of these terms is either 00 or pa=Kp \cdot a = K. If the latter ever happens, the induction statement implies Li=Rn1i=0L_{i} = R_{n-1-i} = 0 together with ai=ak1i=0a_{i} = a_{k-1-i} = 0 and we are finished.

Hence, we can next assume that
K=L0ai+Lia=pai+Lia K = L_{0} a_{i} + L_{i} a = p a_{i} + L_{i} a
and the analogous argument on the other side (replace LjL_{j} by Rn1jR_{n-1-j} and aja_{j} by ak1ja_{k-1-j} ) allows us to assume that
K=Rn1ak1i+Rn1ia K = R_{n-1} a_{k-1-i} + R_{n-1-i} a

Consider the case ai<aa_{i} < a. Equation (1) implies that Li>0L_{i} > 0. Focusing now at the kkth number in John's sequence, we get K=L0ak1++Liak1i+K = L_{0} a_{k-1} + \cdots + L_{i} a_{k-1-i} + \ldots. But L0ak1=pa=KL_{0} a_{k-1} = p a = K, so we are getting Liak1i=0L_{i} a_{k-1-i} = 0, hence ak1i=0a_{k-1-i} = 0.

We can apply exactly the same argument on the other side (replace LjL_{j} with Rn1jR_{n-1-j} and aja_{j} with ak1ja_{k-1-j} and eq. (1) with eq. (2)). We get that whenever ak1i<aa_{k-1-i} < a, it has to be that ai=0a_{i} = 0.

Putting these two arguments together, we conclude that one of the two possibilities is true: Either ai=ak1i=aa_{i} = a_{k-1-i} = a and eqs. (1) and (2) imply Li=Rni=0L_{i} = R_{n-i} = 0. Or ai=ak1i=0a_{i} = a_{k-1-i} = 0 and eqs. (1) and (2) imply Li=Rni=pL_{i} = R_{n-i} = p. In either case, the induction statement is proven for ii and we are done.

Number-theoretic solution. As in the first solution, we denote by LiL_{i} the number of times that a0a_{0} was added to the ii-th position and note that a0a_{0} is maximal among the terms of the sequence. Obviously, we are done if a0=0a_{0} = 0, so let's assume a0>0a_{0} > 0, implying L0>0L_{0} > 0. We normalize the sequence by dividing each term by a0a_{0}, so that a0=1a_{0} = 1 and all numbers on the board are equal to L0L_{0} in the end.

Claim. All terms of the sequence are now rational numbers.

Proof. Assume the contrary. Consider the smallest index ii with aiQa_{i} \notin \mathbb{Q} and look at the ii-th number on the board which in the end of the process takes the value L0NL_{0} \in \mathbb{N}. Since aia_{i} was added to this position L0L_{0} times and all other terms added to it must have been rational by minimality of ii, we get a contradiction.

If there now exists some index jj with 0<aj<10 < a_{j} < 1, write aja_{j} as a reduced fraction and take a prime factor qq of its denominator. In other words, vq(aj)<0=vq(a0)v_{q}\left(a_{j}\right) < 0 = v_{q}\left(a_{0}\right).

We now choose a minimal index tt that minimizes vq(at)v_{q}\left(a_{t}\right) and a minimal index ss that minimizes vq(Ls)v_{q}\left(L_{s}\right). By evaluating the number on the (s+t)(s+t)-th position on the blackboard, we obtain
L0a0=i=0s+tLias+ti L_{0} \cdot a_{0} = \sum_{i=0}^{s+t} L_{i} \cdot a_{s+t-i}
Here, we simply set as+ti=0a_{s+t-i} = 0 if s+tiks+t-i \geq k. Note that the term LsatL_{s} \cdot a_{t} appearing in this sum has strictly smaller vqv_{q} than any other term in the sum, by definition and minimality of ss and tt. Also, since vq(Ls)vq(L0)v_{q}\left(L_{s}\right) \leq v_{q}\left(L_{0}\right) and vq(at)<vq(a0)=0v_{q}\left(a_{t}\right) < v_{q}\left(a_{0}\right) = 0, we have vq(L0a0)>vq(Lsat)v_{q}\left(L_{0} \cdot a_{0}\right) > v_{q}\left(L_{s} \cdot a_{t}\right). Hence,
vq(L0a0)>vq(Lsat)=vq(i=0s+tLias+ti) v_{q}\left(L_{0} \cdot a_{0}\right) > v_{q}\left(L_{s} \cdot a_{t}\right) = v_{q}\left(\sum_{i=0}^{s+t} L_{i} \cdot a_{s+t-i}\right)
contradiction. We conclude that no such jj could have existed, so all terms are either equal to 00 or a0a_{0}.

Solution using Polynomials. By the claim of the previous solution, we see that we may assume all terms of the sequence to be nonnegative integers (just multiply everything by the lcm of the denominators). Again, we note that for all ii, it holds that a0aia_{0} \geq a_{i}.

Now observe the polynomial
P(x)=i=0k1aixi P(x) = \sum_{i=0}^{k-1} a_{i} x^{i}
and interpret the numbers on the board as a polynomial in a similar way, i.e. if the numbers on the board are b0,,bn1b_{0}, \ldots, b_{n-1}, read it as
i=0n1bixi \sum_{i=0}^{n-1} b_{i} x^{i}
The problem states that after each step, the polynomial on the board is increased by xpP(x)x^{p} \cdot P(x) for some pN0p \in \mathbb{N}_{0}. Therefore, the condition can be rewritten as
P(x)Q(x)=Kxn1x1 P(x) \cdot Q(x) = K \cdot \frac{x^{n} - 1}{x - 1}
for some polynomial QZ[x]Q \in \mathbb{Z}[x] and KNK \in \mathbb{N}. It follows that
P(x)=aR(x) P(x) = a \cdot R(x)
where aKa \mid K and R(x)xn1x1R(x) \left\lvert\, \frac{x^{n} - 1}{x - 1}\right.. We see that the constant term of RR must be equal to 11, and so a0=aa_{0} = a. As aaia \mid a_{i} and a0aia_{0} \geq a_{i} for all ii, they are indeed all equal to aa.

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.