Maths Olympiad Prep

Library / /518 of 520

Combinatorics Difficulty 7.8 National olympiad, round 2 Prove it

The rows and columns of a 2n×2n2^{n} \times 2^{n} table are numbered from 0 to 2n12^{n}-1. The cells of the table have been colored with the following property being satisfied: for each 0i,j2n10 \leq i, j \leq 2^{n}-1, the jj th cell in the ii th row and the jj th cell in the ii th column have the same color. Prove that the maximal possible number of colors is 2n2^{n}.

Solution

Throughout the solution we denote the cells of the table by coordinate pairs; (i,j)(i, j) refers to the jj th cell in the ii th row. Consider the directed graph, whose vertices are the cells of the board, and the edges are the arrows (i,j)(j,i+j)(i, j) \rightarrow(j, i+j) for all 0i,j2n10 \leq i, j \leq 2^{n}-1. From each vertex (i,j)(i, j), exactly one edge passes (\left(\right. to (j,i+jmod2n)\left(j, i+j \bmod 2^{n}\right) ); conversely, to each cell (j,k)(j, k) exactly one edge is directed (from the cell (kjmod2n,j))\left.\left(k-j \bmod 2^{n}, j\right)\right). Hence, the graph splits into cycles. Now, in any coloring considered, the vertices of each cycle should have the same color by the problem condition. On the other hand, if each cycle has its own color, the obtained coloring obviously satisfies the problem conditions. Thus, the maximal possible number of colors is the same as the number of cycles, and we have to prove that this number is 2n2^{n}. Next, consider any cycle (i1,j1),(i2,j2),\left(i_{1}, j_{1}\right),\left(i_{2}, j_{2}\right), \ldots; we will describe it in other terms. Define a sequence (a0,a1,)\left(a_{0}, a_{1}, \ldots\right) by the relations a0=i1,a1=j1,an+1=an+an1a_{0}=i_{1}, a_{1}=j_{1}, a_{n+1}=a_{n}+a_{n-1} for all n1n \geq 1 (we say that such a sequence is a Fibonacci-type sequence). Then an obvious induction shows that ikak1(mod2n),jkak(mod2n)i_{k} \equiv a_{k-1}\left(\bmod 2^{n}\right), j_{k} \equiv a_{k}\left(\bmod 2^{n}\right). Hence we need to investigate the behavior of Fibonacci-type sequences modulo 2n2^{n}. Denote by F0,F1,F_{0}, F_{1}, \ldots the Fibonacci numbers defined by F0=0,F1=1F_{0}=0, F_{1}=1, and Fn+2=F_{n+2}= Fn+1+FnF_{n+1}+F_{n} for n0n \geq 0. We also set F1=1F_{-1}=1 according to the recurrence relation. For every positive integer mm, denote by ν(m)\nu(m) the exponent of 2 in the prime factorization of mm, i.e. for which 2ν(m)m2^{\nu(m)} \mid m but 2ν(m)+1\m2^{\nu(m)+1} \backslash m. Lemma 1. For every Fibonacci-type sequence a0,a1,a2,a_{0}, a_{1}, a_{2}, \ldots, and every k0k \geq 0, we have ak=a_{k}= Fk1a0+Fka1F_{k-1} a_{0}+F_{k} a_{1}. Proof. Apply induction on kk. The base cases k=0,1k=0,1 are trivial. For the step, from the induction hypothesis we get
ak+1=ak+ak1=(Fk1a0+Fka1)+(Fk2a0+Fk1a1)=Fka0+Fk+1a1 a_{k+1}=a_{k}+a_{k-1}=\left(F_{k-1} a_{0}+F_{k} a_{1}\right)+\left(F_{k-2} a_{0}+F_{k-1} a_{1}\right)=F_{k} a_{0}+F_{k+1} a_{1}
Lemma 2. For every m3m \geq 3, (a) we have ν(F32m2)=m\nu\left(F_{3 \cdot 2^{m-2}}\right)=m; (b) d=32m2d=3 \cdot 2^{m-2} is the least positive index for which 2mFd2^{m} \mid F_{d}; (c) F32m2+11+2m1(mod2m)F_{3 \cdot 2^{m-2}+1} \equiv 1+2^{m-1}\left(\bmod 2^{m}\right). Proof. Apply induction on mm. In the base case m=3m=3 we have ν(F32m2)=F6=8\nu\left(F_{3 \cdot 2^{m-2}}\right)=F_{6}=8, so ν(F32m2)=ν(8)=3\nu\left(F_{3 \cdot 2^{m-2}}\right)=\nu(8)=3, the preceding Fibonacci-numbers are not divisible by 8 , and indeed F32m2+1=F7=131+4(mod8)F_{3 \cdot 2^{m-2}+1}=F_{7}=13 \equiv 1+4(\bmod 8). Now suppose that m>3m>3 and let k=32m3k=3 \cdot 2^{m-3}. By applying Lemma 1 to the Fibonacci-type sequence Fk,Fk+1,F_{k}, F_{k+1}, \ldots we get
F2k=Fk1Fk+FkFk+1=(Fk+1Fk)Fk+Fk+1Fk=2Fk+1FkFk2F2k+1=FkFk+Fk+1Fk+1=Fk2+Fk+12 \begin{gathered} F_{2 k}=F_{k-1} F_{k}+F_{k} F_{k+1}=\left(F_{k+1}-F_{k}\right) F_{k}+F_{k+1} F_{k}=2 F_{k+1} F_{k}-F_{k}^{2} \\ F_{2 k+1}=F_{k} \cdot F_{k}+F_{k+1} \cdot F_{k+1}=F_{k}^{2}+F_{k+1}^{2} \end{gathered}
By the induction hypothesis, ν(Fk)=m1\nu\left(F_{k}\right)=m-1, and Fk+1F_{k+1} is odd. Therefore we get ν(Fk2)=\nu\left(F_{k}^{2}\right)= 2(m1)>(m1)+1=ν(2FkFk+1)2(m-1)>(m-1)+1=\nu\left(2 F_{k} F_{k+1}\right), which implies ν(F2k)=m\nu\left(F_{2 k}\right)=m, establishing statement (a). Moreover, since Fk+1=1+2m2+a2m1F_{k+1}=1+2^{m-2}+a 2^{m-1} for some integer aa, we get
F_{2 k+1}=F_{k}^{2}+F_{k+1}^{2} \equiv 0+\left(

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: NuminaMath-1.5, licensed Apache-2.0. Statement and solution reproduced as published; topic and difficulty added by this site.