Maths Olympiad Prep

Track / Stage 7 / 267 of 300 #1667 of 1964

Problem 1667

National Olympiad second round; IMO P1/P4
Number theory Difficulty 7.6 Prove it

8-51 Let the sequence of integers {an}\left\{a_{n}\right\} satisfy:
a0=0,a1=1,an=2an−1+an−2,n=2,3,4,⋯a_{0}=0, a_{1}=1, a_{n}=2 a_{n-1}+a_{n-2}, n=2,3,4, \cdots

Prove: 2k∣an⇔2k∣n,k=0,1,2,⋯2^{k}\left|a_{n} \Leftrightarrow 2^{k}\right| n, k=0,1,2, \cdots

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

[Proof] It is sufficient to prove the case for k⩾1k \geqslant 1. From the recursive formula, we have
an≡an−2( mod 2),n=2,3,4,⋯a_{n} \equiv a_{n-2}(\bmod 2), n=2,3,4, \cdots

Given a0=0,a1=1a_{0}=0, a_{1}=1, we have
2∣an⇔2∣n,2\left|a_{n} \Leftrightarrow 2\right| n,

which means the conclusion holds for k=1k=1. Next, we prove the case for k>1k>1. From the recursive formula and a0=0,a1=1a_{0}=0, a_{1}=1, we can easily obtain the general term formula
an=24[(1+2)n−(1−2)n],a_{n}=\frac{\sqrt{2}}{4}\left[(1+\sqrt{2})^{n}-(1-\sqrt{2})^{n}\right],

Thus, a2mn=24[(1+2)2′′′n−(1−222mn]\quad a_{2^{m} n}=\frac{\sqrt{2}}{4}\left[(1+\sqrt{2})^{2^{\prime \prime \prime} n}-\left(1-\sqrt{2}^{2^{2 m} n}\right]\right..
Let (1+2)2m+(1−2)2m=2∑c2km2k2=2b2m′\quad(1+\sqrt{2})^{2 m}+(1-\sqrt{2})^{2^{m}}=2 \sum c_{2}^{k m} 2^{\frac{k}{2}}=2 b_{2}{ }^{m \prime},
where the sum is over all even numbers in [0,2m]\left[0,2^{m}\right], and
(1+2)2′′′⋅(1−2)2m={−1,m=01,m>0(1+\sqrt{2})^{2^{\prime \prime \prime}} \cdot(1-\sqrt{2})^{2^{m}}=\left\{\begin{array}{r} -1, m=0 \\ 1, m>0 \end{array}\right.

Therefore, [λ−(1+2)2m][λ−(1−2)2′′]={λ2−2b2mλ−1,m=0,λ2−2b2mλ+1,m>0.\left[\lambda-(1+\sqrt{2})^{2^{m}}\right]\left[\lambda-(1-\sqrt{2})^{2^{\prime \prime}}\right]=\left\{\begin{array}{l}\lambda^{2}-2 b_{2}{ }^{m} \lambda-1, m=0, \\ \lambda^{2}-2 b_{2}{ }^{m} \lambda+1, m>0 .\end{array}\right. From this, we can derive the recursive formula for the sequence {a2mn},n=0,1,2,⋯\left\{a_{2}{ }^{m}{ }_{n}\right\}, n=0,1,2, \cdots as
a2mn={2b2ma2′′′(n−1)+a2′′′(n−2),m=0,2b2′′a2m′′(n−1)−a2m′(n−2),m>0.a_{2}{ }^{m}{ }_{n}=\left\{\begin{array}{l} 2 b_{2}{ }^{m} a_{2}{ }^{\prime \prime \prime}(n-1)+a_{2}{ }^{\prime \prime \prime}(n-2), m=0, \\ 2 b_{2}{ }^{\prime \prime} a_{2}{ }^{m \prime \prime}(n-1)-a_{2}{ }^{m \prime}(n-2), m>0 . \end{array}\right.

Consider the sequence bn=12[(1+2)n+(1−2)n],n=0,1,2,⋯\quad b_{n}=\frac{1}{2}\left[(1+\sqrt{2})^{n}+(1-\sqrt{2})^{n}\right], n=0,1,2, \cdots
It is easy to see that b0=b1=1,bn=2bn−1+bn−2b_{0}=b_{1}=1, b_{n}=2 b_{n-1}+b_{n-2}, for any n⩾2n \geqslant 2.
Thus, for any non-negative integer n,bnn, b_{n} is odd. From the recursive formula, we can see that when m⩾1m \geqslant 1, we have

Thus, a2m=2mb2m−1b2mm⋯b1\quad a_{2}{ }^{m}=2^{m} b_{2^{m-1}} b_{2^{m}}{ }^{m} \cdots b_{1}.
Since b1,b2,⋯ ,b2m−1b_{1}, b_{2}, \cdots, b_{2}{ }^{m-1} are all odd, we have
2m∣∣a2m2^{m}|| a_{2^{m}}

which means 2m∣a2m2^{m} \mid a_{2}{ }^{m} but 2m+1∣a2m2^{m+1} \mid a_{2}{ }^{m}. From the recursive formula, we can also get
2m∣a2mn and a2m(n+2)≡a2mn( mod 2m+1),n=0,1,2⋯ . 2^{m} \mid a_{2^{m}}{ }_{n} \text { and } a_{2}{ }^{m}(n+2) \equiv a_{2} m_{n}\left(\bmod 2^{m+1}\right), n=0,1,2 \cdots \text { . }

From 2m+1∣a0=0,2m+1×a2m2^{m+1} \mid a_{0}=0,2^{m+1} \times a_{2} m, we can conclude that for any m⩾1m \geqslant 1,
2m+1∣a2mn⇔2∣n.2^{m+1}\left|a_{2^{m}}{ }_{n} \Leftrightarrow 2\right| n .

For any natural number k>1k>1. If 2k∣n2^{k} \mid n, then n=2k−1ln=2^{k-1} l, where ll is even. From the above result, we have 2k∣an2^{k} \mid a_{n}. Conversely, if 2k×n2^{k} \times n, then there exists 0⩽m⩽k−10 \leqslant m \leqslant k-1 such that n=2mln=2^{m} l, where ll is odd. Thus, 2m+1×an2^{m+1} \times a_{n}, and since k⩾m+1k \geqslant m+1, we have 2k×an2^{k} \times a_{n}. The above shows that for k>1k>1, the conclusion also holds.

Source: NuminaMath-1.5, licensed Apache-2.0. Statement reproduced verbatim; metadata (topic, difficulty, ordering) added by this project.