Maths Olympiad Prep

Library / /114 of 120

, 2012

Number theory Difficulty 6.6 National olympiad Prove it Saudi Arabia

Let pp be a prime. At any vertex of a regular polygon with pp sides it is written an integer. For any vertex of the polygon we compute the difference between the sum of the integers written at his neighbors and his number. After that we delete all the initial integers and replace them by the new obtained integers. Prove that the integers obtained after pp such steps are the same modulo pp with the initial integers.

Solutions — 2

Solution 1

Let A0A1...Ap1A_0A_1...A_{p-1} be the regular polygon and let (a0,a1,...,ap1)(a_0, a_1, ..., a_{p-1}) be the integers written at its vertices, aka_k at AkA_k for k=0,1,...,p1k = 0, 1, ..., p-1. Consider the polynomial with integer coefficients
H(x)=a0+a1x+...+ap1xp1. H(x) = a_0 + a_1x + ... + a_{p-1}x^{p-1}.
Applying the transformation
a0ap1+a1a0,a1a0+a2a1,,ap1ap2+a0ap1 a_0 \to a_{p-1} + a_1 - a_0, \quad a_1 \to a_0 + a_2 - a_1, \dots, a_{p-1} \to a_{p-2} + a_0 - a_{p-1}
on the coefficients of the polynomial H(x)H(x) we obtain the polynomial
H1(x)=H(x)(xp1+x1)(mod(xp1)). H_1(x) = H(x)(x^{p-1} + x - 1) \pmod{(x^p - 1)}.
After ss steps we obtain the polynomial
Hs(x)=Hs1(x)(xp1+x1)(mod(xp1))=H(x)(xp1+x1)s(mod(xp1)). \begin{aligned} H_s(x) &= H_{s-1}(x)(x^{p-1} + x - 1) \pmod{(x^p - 1)} \\ &= H(x)(x^{p-1} + x - 1)^s \pmod{(x^p - 1)}. \end{aligned}
Finally, after pp steps we get
Hp(x)=(xp1+x1)pH(x)(mod(xp1))(1) H_p(x) = (x^{p-1} + x - 1)^p H(x) \pmod{(x^p - 1)} \quad (1)
But
(xp1+x1)p=pQ(x)+xp(p1)+xp1 (x^{p-1} + x - 1)^p = p \cdot Q(x) + x^{p(p-1)} + x^p - 1
=pQ(x)+(xp)p11+xp1+1=pQ(x)+1(mod(xp1))= p \cdot Q(x) + (x^p)^{p-1} - 1 + x^p - 1 + 1 = p \cdot Q(x) + 1 \pmod{(x^p - 1)},
where QQ is a polynomial with integer coefficients. From (1) we get
Hp(x)=pQ(x)H(x)+H(x)(mod(xp1)), H_p(x) = p \cdot Q(x)H(x) + H(x) \pmod{(x^{p-1})},
Therefore HpH_p and HH have the same coefficients modulo pp.

Solution 2

The transformation in the problem is described by the following matrix relation
(a0(1)ap1(1))=(11001111000110010011)(a0ap1)(1) \begin{pmatrix} a_0^{(1)} \\ \vdots \\ a_{p-1}^{(1)} \end{pmatrix} = \begin{pmatrix} -1 & 1 & 0 & \cdots & 0 & 1 \\ 1 & -1 & 1 & \cdots & 0 & 0 \\ 0 & 1 & -1 & \cdots & 0 & 0 \\ \vdots & \vdots & \vdots & \ddots & \vdots & \vdots \\ 1 & 0 & 0 & \cdots & 1 & -1 \end{pmatrix} \begin{pmatrix} a_0 \\ \vdots \\ a_{p-1} \end{pmatrix} \quad (1)
After pp steps we get
(a0(p)ap1(p))=Ap(a0ap1),(2) \begin{pmatrix} a_0^{(p)} \\ \vdots \\ a_{p-1}^{(p)} \end{pmatrix} = A^p \cdot \begin{pmatrix} a_0 \\ \vdots \\ a_{p-1} \end{pmatrix}, \quad (2)
where AA is the square matrix in relation (1).
We can write A=Ip+X+Xp1A = -I_p + X + X^{p-1}, where XX is the permutation matrix
X=(0001100001000010). X = \begin{pmatrix} 0 & 0 & \dots & 0 & 1 \\ 1 & 0 & \dots & 0 & 0 \\ 0 & 1 & \dots & 0 & 0 \\ \dots & \dots & \dots & \dots & \dots \\ 0 & 0 & \dots & 1 & 0 \end{pmatrix} .
In Mp(Zp)M_p(\mathbb{Z}_p) we have
Ap=(Ip+X+Xp1)p=(Ip)p+Xp+Xp(p1)=Ip+Ip+Ip=Ip, \begin{align*} A^p &= (-I_p + X + X^{p-1})^p = (-I_p)^p + X^p + X^{p(p-1)} \\ &= -I_p + I_p + I_p = I_p, \end{align*}
since Xp=IpX^p = I_p. From (2) it follows that in Mp,1(Zp)M_{p,1}(\mathbb{Z}_p) we have
(a0(p)ap1(p))=(a0ap1), \begin{pmatrix} a_0^{(p)} \\ \vdots \\ a_{p-1}^{(p)} \end{pmatrix} = \begin{pmatrix} a_0 \\ \vdots \\ a_{p-1} \end{pmatrix},
and we are done.

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 and solution reproduced as published; topic and difficulty added by this site.