奥数自学研究

高中奥数 2022-01-25

2022-01-25  本文已影响0人  不为竞赛学奥数

2022-01-25-01

(来源: 数学奥林匹克小丛书 第二版 高中卷 数列与数学归纳法 冯志刚 习题一 P055 习题41)

数列\left\{y_{n}\right\}定义如y_{2}=y_{3}=1,且

\left(n+1\right)\left(n-2\right)y_{n+1}=n\left(n^{2}-n-1\right)y_{n}-\left(n-1\right)^{3}y_{n-1}.,n=3,4,\cdots

证明:y_{n}为整数的充要条件是n为素数.

证明

x_{n}=ny_{n},n=2,3,\cdots ,则x_{2}=2,x_{3}=3,且当n\geqslant 3时,有\left(n-2\right)x_{n+1}=\left(n^{2}-n-1\right)x_{n}-\left(n-1\right)^{2}x_{n-1},即

\dfrac{x_{n-1}-x_{n}}{n-1}=\left(n-1\right)\cdot\dfrac{x_{n}-x_{n-1}}{n-2}.\qquad(1)

利用(1)递推可得

\begin{aligned} \dfrac{x_{n+1}-x_{n}}{n-1} &=(n-1) \cdot \dfrac{x_{n}-x_{n-1}}{n-2}\\ &=(n-1)(n-2) \cdot \dfrac{x_{n-1}-x_{n-2}}{n-3} \\ &=\cdots\\ &=(n-1) \cdots 2 \cdot \dfrac{x_{3}-x_{2}}{1}\\ &=(n-1) ! \end{aligned}

x_{n+1}-x_{n}=n!-\left(n-1\right)!,裂项求和,知x_{n}=x_{2}+\sum\limits_{k=2}^{n-1}\left(x_{k+1}-x_{k}\right)=x_{2}+\sum\limits_{k=2}^{n-1}\left(k!-\left(k-1\right)!\right)=x_{2}+\left(n-1\right)!-1=\left(n-1\right)!+1.

结合Wilson定理:当且仅当n为素数时,n\mid \left(n-1\right)!+1,可知y_{n}\in \mathbb{Z}的充要条件是n为素数.

2022-01-25-02

(来源: 数学奥林匹克小丛书 第二版 高中卷 数列与数学归纳法 冯志刚 习题一 P056 习题42)

给定素数p>3,令q=p^{3}.定义数列a_{n}如下

a_{n}=\begin{cases} n,&n=0,1,2,\cdots,p-1,\\ a_{n-1}+a_{n-p},&n>p-1. \end{cases}.

a_{q}除以p所得的余数.

引理nk\in \mathbb{N}^{*},n\geqslant kp,则

a_{n}=\sum\limits_{i=0}^{k}\mathrm{C}_{k}^{i}a_{n-i\left(p-1\right)-k}.\qquad(1)

k归纳予以证明.当k=1时,(1)就是a_{n}=a_{n-1}+a_{n-p},故(1)对k=1成立.

现设(1)对k成立,考虑k+1的情形.此时n\geqslant \left(k+1\right)p,下标n-i\left(p-1\right)-k\left(0\leqslant i\leqslant k\right)的最小值在i=k时取到,该最小值为n-kp\geqslant p,所以,下面求和式中的每一项都可用条件中的递推式.

由归纳假设知,当n\geqslant \left(k+1\right)p时,有

\begin{aligned} a_{n} &=\sum\limits_{i=0}^{k} \mathrm{C}_{k}^{i} a_{n-i(p-1)-k} \\ &=\sum\limits_{i=0}^{k} \mathrm{C}_{k}^{i}\left(a_{n-i(p-1)-k-1}+a_{n-i(p-1)-k-p}\right) \\ &=\mathrm{C}_{k}^{0} a_{n-k-1}+\sum\limits_{i=1}^{k} \mathrm{C}_{k}^{i} a_{n-i(p-1)-k-1}+\sum\limits_{i=0}^{k-1} \mathrm{C}_{k}^{i} a_{n-i(p-1)-k-p}+\mathrm{C}_{k}^{k} a_{n-(k+1) p} \\ &=\mathrm{C}_{k+1}^{0} a_{n-(k+1)}+\sum\limits_{i=0}^{k-1} \mathrm{C}_{k}^{i+1} a_{n-(i+1)(p-1)-(k+1)}+\sum\limits_{i=0}^{k-1} \mathrm{C}_{k}^{i} a_{n-(i+1)(p-1)-(k+1)}+\mathrm{C}_{k+1}^{k+1} a_{n-(k+1) p} \\ &=\mathrm{C}_{k+1}^{0} a_{n-(k+1)}+\sum\limits_{i=0}^{k-1}\left(\mathrm{C}_{k}^{i+1}+\mathrm{C}_{k}^{i}\right) a_{n-(i+1)(p-1)-(k+1)}+\mathrm{C}_{k+1}^{k+1} a_{n-(k+1) p} \\ &=\sum\limits_{i=0}^{k+1} \mathrm{C}_{k+1}^{i} a_{n-(i+1)(p-1)-(k+1)} . \end{aligned}

最后一步,用到\mathrm{C}_{k}^{i+1}+\mathrm{C}_{k}^{i}=\mathrm{C}_{k+1}^{i+1}.

所以,(1)对k+1成立,引理获证.

下面利用引理来处理原题.

n\geqslant p^{2}时,在引理中令k=p,就有

a_{n}=\sum\limits_{i=0}^{p} \mathrm{C}_{p}^{i} a_{n-i(p-1)-p}

熟知,当1\leqslant i\leqslant p-1时,有\mathrm{C}_{p}^{i}\equiv 0\left(\bmod p\right),所以,a_{n}\equiv a_{n-p}+a_{n-p^{2}}\left(\bmod p\right),这时结合a_{n}=a_{n-1}+a_{n-p},可得a_{n-1}\equiv a_{n-p^{2}}\left(\bmod p\right),这表明对任意t\geqslant p^{2}-1,有a_{t}\equiv a_{t+p^{2}-1}\left(\bmod p\right).

由于p^{3}=p\left(p^{2}-1\right)+p,故a_{p^{3}}=a_{p+p\left(p^{2}-1\right)}\equiv a_{p}\left(\bmod p\right),而a_{p}=a_{0}+a_{p-1}=p-1,所以a_{p^{3}}\equiv p-1\left(\bmod p\right),即a_{p^3}除以p所得的余数为p-1.

2022-01-25-03

(来源: 数学奥林匹克小丛书 第二版 高中卷 数列与数学归纳法 冯志刚 习题一 P056 习题43)

n为不小于2的正整数,2\leqslant b_{0}\leqslant 2n-1,b_{0}为整数,考虑由递推式

\b_{i+1}=\begin{cases} 2b_{i}-1,&\text{若}b_{i}\leqslant n,\\ 2b_{i}-2n,&\text{若}b_{i}>n \end{cases}

定义的数列\left\{b_{i}\right\}.用p\left(b_{0},n\right)表示满足b_{p}=b_{0}的最小下标p.

(i)设k为给定的正整数,求p\left(2,2^{k}\right)p\left(2,2^{k}+1\right)的值;

(ii)证明:对任意nb_{0}都有p\left(b_{0},n\right)\mid p\left(2,n\right).

为方便计,记m=n-1,b_{i}=a_{i}+1,则1\leqslant a_{0}\leqslant 2m,且

a_{i+1}=\begin{cases} 2a_{i},&\text{若}a_{i}\leqslant m,\\ 2a_{i}-\left(2m+1\right),&\text{若}a_{i}>m. \end{cases}

这表明a_{i+1}\equiv 2a_{i}\left(\bmod 2m+1\right),且1\leqslant a_{i}\leqslant 2m,i=1,2,\cdots.

(i)题中的p\left(2,2^{k}\right)p\left(2,2^{k}+1\right)等价于针对\left\{a_{i}\right\}p\left(1,2^{k}-1\right)p\left(1,2^{k}\right).

前者等价于求最小l\in \mathbb{N}^{*},使得2^{l}\equiv 1\left(\bmod 2\left(2^{k}-1\right)+1\right),后者等价于求最小的l\in \mathbb{N}^{*},使得2^{l}\equiv 1\left(\bmod 2^{k+1}+1\right).

显然2^{k+1}\equiv 1\left(\bmod 2\left(2^{k}-1\right)+1\right),而对1\leqslant t\leqslant k,都有1\leqslant 2^{t}-1<2^{k-1}-1=2\left(2^{k}-1\right)+1,故p\left(1,2^{k}-1\right)=k+1.

2^{2\left(k+1\right)}\equiv 1\left(\bmod2^{k+1}+1\right),从而p\left(1,2^{k}\right)\mid 2\left(k+1\right),又对1\leqslant t\leqslant k+1,都有1\leqslant 2^{t}-1<2^{k+1}+1,于是p\left(1,2^{k}\right)>k+1,故p\left(1,2^{k}\right)=2\left(k+1\right).

所以,针对\left\{b_{i}\right\}p\left(2,2^{k}\right)=k+1,p\left(2,2^{k}+1\right)=2\left(k+1\right).

(ii)还是转到\left\{a_{i}\right\}上讨论,要求证明:\left(a_{0},m\right)\mid p\left(1,m\right).

现设p\left(1,m\right)=t,则2^{t}\equiv 1\left(\bmod 2m+1\right),从而2^{t}a_{0}\equiv a_{0}\left(\bmod 2m+1\right),这表明p\left(a_{0},m\right)\mid t(这里用到类似于初等数论中阶的性质),即有p\left(a_{0},m\right)\mid p\left(1,m\right),命题成立.

上一篇 下一篇

猜你喜欢

热点阅读