Expression is a perfect square implies the polynomial is constant

Art of Problem Solving

AoPS Online

High School Olympiads Regional, national, and international math olympiads

Regional, national, and international math olympiads

3 M G

BBookmark VNew Topic kLocked

High School Olympiads Regional, national, and international math olympiads

Regional, national, and international math olympiads

3 M G

BBookmark VNew Topic kLocked

No tags match your search

M

Expression is a perfect square implies the polynomial is constant

G H J

Reveal topic tags

number theory proposed

algebra proposed

Lifting the Exponent

L

G H BBookmark kLocked kLocked NReply

Source: IMSC Day 2 Problem 3

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

109 posts

#1 h Jun 29, 2024, 8:08 AM • 1 Y

Y by MathLuis

Let $a\equiv 1\pmod{4}$ be a positive integer. Show that any polynomial $Q\in\mathbb{Z}[X]$ with all positive coefficients such that
$Q(n+1)((a+1)^{Q(n)}-a^{Q(n)})$ is a perfect square for any $n\in\mathbb{N}^{\ast}$ must be a constant polynomial.

Proposed by Vlad Matei, Romania

This post has been edited 2 times. Last edited by Popescu, Jun 30, 2024, 12:36 PM

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

40 posts

#2 h Jul 5, 2024, 4:31 AM

Y by

The idea is quite straightforward but the details are a little bit cumbersome.I'll just show the sketch of my solution.If $Q(x)$ is not in the form of $cP(x)^2$ ,where $c\in\mathbb{N^*}$ and $P\in\mathbb{Z}[x]$ ,we could find a sufficiently large prime number $p$ satisfied that $p\nmid (a+1)^{Q(1)}-a^{Q(1)}$ and $p\mid Q_{1}(n_{p}+1)$ ,but $p\nmid Q'_{1}(n_{p}+1)$ for some $n_p\in\mathbb{N^*}$ ,where $Q_{1}(x)$ is the squarefree part of $Q(x)$ which has degree $d\geq 1$ .Then there exist $\alpha\in\mathbb{N^*}$ and $p^{2\alpha}\Vert Q_{1}(n_{p}+1)$ .By Hensel Lemma and CRT we could find $n\equiv n_p\mod p^{2\alpha}$ and $n\equiv 1\mod p-1$ satisfied that $p^{2\alpha+1}\Vert Q_{1}(n+1)$ and $p\nmid (a+1)^{Q(n)}-a^{Q(n)}$ ,which gives a contradition.If $Q(x)=cP(x)^2$ ,then $c((a+1)^{Q(n)}-a^{Q(n)})$ is also a perfect square.Choose prime $q$ sufficiently large so that $\gcd (c,(a+1)^q-a^q)=1$ and $q\mid Q(m)$ for some $m\in\mathbb{N^*}$ .Since $(a+1)^q-a^q$ isn't a perfect square,we could choose a prime factor $p$ satisfied that $p^{2\alpha-1}\Vert (a+1)^q-a^q$ for some $\alpha\in\mathbb{N^*}$ .It's well-known that $q\mid p-1$ so $p$ is sufficiently large,hence there exist $s\in\mathbb{N^*}$ and $p\nmid Q(s)$ .By CRT we could choose $n\equiv m\mod p-1$ and $n\equiv s\mod p$ ,so that $q\mid Q(n)$ and by LTE Lemma, $p^{2\alpha-1}\Vert (a+1)^{Q(n)}-a^{Q(n)}$ ,a contradiction.And the problem is done.

Z K Y

The post below has been deleted. Click to close.

This post has been deleted. Click here to see post.

661 posts

#4 h Jan 11, 2025, 9:18 PM • 1 Y

Y by MS_asdfgzxcvb

Solved with erkosfobiladol.
$Q(n+1)((a+1)^{Q(n)}-a^{Q(n)})\in \mathbb{Z}^2$ Claim: There are finitely many primes $p$ such that there exists $v_p(Q(l))\equiv 1(mod \ 2)$ .
Proof: We will show that $v_p(Q(l))\equiv 1(mod \ 2)$ implies $p|(a+1)^{Q(1)}-a^{Q(1)}$ . Suppose that $p\not | (a+1)^{Q(1)}-a^{Q(1)}$ and $p^{2m+1}||Q(n+1)$ . We see that $v_p(Q(n+p^{2m+2}k+1))=2m+1$ hence $p|(a+1)^{Q(n+p^{2m+2}k+1)}-a^{Q(n+p^{2m+2}k+1}$ or $p|(a+1)^{Q(n+k+1)}-a^{Q(n+k+1)}$ . Pick $k\equiv -n(mod \ p-1)$ in order to get a contradiction. $\square$
Claim: $Q(x)=cR(x)^2$ for some $R\in \mathbb{Z}[X]$ and a squarefree constant $c$ .
Proof: Let $Q(x)=cA_1(x)^{t_1}\dots A_n(x)^{t_n}$ where $A_i'$ s are irreducible over integers. By Bezout, let $U_{i,j}A_i+V_{i,j}A_j=C_{i,j}$ and $C=lcm(C_{i,j})$ . Pick a prime $p|A_i(x)$ and $p\not | C$ which exists by Schur. Since $(A_i(x),A_i'(x))=1$ , by "extended" Hensel (we utilise $A_i$ is irreducible), we can choose $v_p(A_i(x))\equiv 1(mod \ 2)$ and this yields $t_i$ is even. Applying this process for each $i$ gives the result. $\square$
$c((a+1)^{cR(n)^2}-a^{cR(n)^2})\in \mathbb{Z}^2$ Suppose that $deg \ R\geq 1$ . Notice that $(a+1)^{R(n)^2}-a^{R(n)^2}\equiv-1(mod \ 4)$ thus, it cannot be a perfect square.
Let $v_p((a+1)^{R(n)^2}-a^{R(n)^2})$ be odd. By LTE, we observe the following equations,
$v_p(c((a+1)^{cR(n)^2}-a^{cR(n)^2}))=v_p(c)+v_p((a+1)^{cR(n)^2}-a^{cR(n)^2})=2v_p(c)+v_p((a+1)^{R(n)^2}-a^{R(n)^2})$ However, $2v_p(c)+v_p((a+1)^{R(n)^2}-a^{R(n)^2})$ is odd which gives a contradiction as desired. $\blacksquare$

Z K Y

N Quick Reply

G

H

=

© 2025 AoPS Incorporated

a