Difference between revisions of "2025 USAMO Problems/Problem 2"
(→Solution) |
(→Solution) |
||
Line 2: | Line 2: | ||
Let <math>n</math> and <math>k</math> be positive integers with <math>k<n</math>. Let <math>P(x)</math> be a polynomial of degree <math>n</math> with real coefficients, nonzero constant term, and no repeated roots. Suppose that for any real numbers <math>a_0, a_1, \dots, a_k</math> such that the polynomial <math>a_kx^k+\cdots+a_1x+a_0</math> divides <math>P(x)</math>, the product <math>a_0a_1\cdots a_k</math> is zero. Prove that <math>P(x)</math> has a nonreal root. | Let <math>n</math> and <math>k</math> be positive integers with <math>k<n</math>. Let <math>P(x)</math> be a polynomial of degree <math>n</math> with real coefficients, nonzero constant term, and no repeated roots. Suppose that for any real numbers <math>a_0, a_1, \dots, a_k</math> such that the polynomial <math>a_kx^k+\cdots+a_1x+a_0</math> divides <math>P(x)</math>, the product <math>a_0a_1\cdots a_k</math> is zero. Prove that <math>P(x)</math> has a nonreal root. | ||
− | == | + | == Problem == |
− | + | Let <math>n</math> and <math>k</math> be positive integers with <math>k<n</math>. Let <math>P(x)</math> be a polynomial of degree <math>n</math> with real coefficients, nonzero constant term, and no repeated roots. Suppose that for any real numbers <math>a_0, a_1, \dots, a_k</math> such that the polynomial <math>a_kx^k+\cdots+a_1x+a_0</math> divides <math>P(x)</math>, the product <math>a_0a_1\cdots a_k</math> is zero. Prove that <math>P(x)</math> has a nonreal root. | |
− | + | == Solution == | |
− | + | We proceed by contradiction. Assume that all roots of <math>P(x)</math> are real. Let the distinct roots be <math>r_1, r_2, \ldots, r_n</math>, all nonzero since the constant term is nonzero. | |
− | |||
− | |||
− | |||
− | |||
− | + | Consider any subset of <math>k</math> roots <math>\{r_{i_1}, r_{i_2}, \ldots, r_{i_k}\}</math> and form the polynomial: | |
− | <cmath> | + | <cmath> Q(x) = \prod_{j=1}^k (x - r_{i_j}) = x^k + a_{k-1}x^{k-1} + \cdots + a_0 </cmath> |
− | |||
− | |||
− | |||
− | \ | + | By Vieta's formulas: |
− | + | \begin{itemize} | |
− | < | + | \item <math>a_0 = (-1)^k \prod_{j=1}^k r_{i_j} \neq 0</math> |
− | + | \item <math>a_{k-1} = -\sum_{j=1}^k r_{i_j}</math> | |
− | < | + | \item <math>a_{k-2} = \sum_{1\leq m<n\leq k} r_{i_m}r_{i_n}</math> |
− | + | \end{itemize} | |
− | Thus, <math>P(x)</math> must have at least one nonreal root. | + | The given condition requires that <math>a_0a_1\cdots a_k = 0</math>. Since <math>a_0 \neq 0</math>, at least one other coefficient must be zero. |
+ | |||
+ | Case <math>k=2</math>: | ||
+ | For any pair of roots <math>(r_i, r_j)</math>, we have: | ||
+ | <cmath> Q(x) = x^2 - (r_i+r_j)x + r_ir_j </cmath> | ||
+ | The condition implies <math>-r_ir_j(r_i+r_j) = 0</math>, so <math>r_i + r_j = 0</math> for all pairs. But with <math>n \geq 3</math>, considering three roots <math>r_1, r_2, r_3</math> gives: | ||
+ | <cmath> r_1 + r_2 = 0 \quad \text{and} \quad r_1 + r_3 = 0 \implies r_2 = r_3 </cmath> | ||
+ | contradicting distinct roots. | ||
+ | In General <math>k</math>: | ||
+ | For any <math>k</math> roots, some symmetric sum must be zero. For <math>k=3</math>, this would require: | ||
+ | <cmath> r_i + r_j + r_m = 0 \quad \text{for all triples} </cmath> | ||
+ | which leads to contradictions when <math>n \geq 4</math> as it would force roots to be equal. | ||
+ | |||
+ | Thus, our initial assumption is false, and <math>P(x)</math> must have at least one nonreal root.(By Jonathan) | ||
+ | |||
+ | == See Also == | ||
+ | {{USAMO newbox|year=2025|num-b=1|num-a=3}} | ||
+ | {{WAA Notice}} | ||
== See Also == | == See Also == | ||
{{USAMO newbox|year=2025|num-b=1|num-a=3}} | {{USAMO newbox|year=2025|num-b=1|num-a=3}} | ||
{{WAA Notice}} | {{WAA Notice}} |
Revision as of 23:47, 27 March 2025
Contents
Problem
Let and
be positive integers with
. Let
be a polynomial of degree
with real coefficients, nonzero constant term, and no repeated roots. Suppose that for any real numbers
such that the polynomial
divides
, the product
is zero. Prove that
has a nonreal root.
Problem
Let and
be positive integers with
. Let
be a polynomial of degree
with real coefficients, nonzero constant term, and no repeated roots. Suppose that for any real numbers
such that the polynomial
divides
, the product
is zero. Prove that
has a nonreal root.
Solution
We proceed by contradiction. Assume that all roots of are real. Let the distinct roots be
, all nonzero since the constant term is nonzero.
Consider any subset of roots
and form the polynomial:
By Vieta's formulas:
\begin{itemize}
\item
\item
\item
\end{itemize}
The given condition requires that . Since
, at least one other coefficient must be zero.
Case :
For any pair of roots
, we have:
The condition implies
, so
for all pairs. But with
, considering three roots
gives:
contradicting distinct roots.
In General
:
For any
roots, some symmetric sum must be zero. For
, this would require:
which leads to contradictions when
as it would force roots to be equal.
Thus, our initial assumption is false, and must have at least one nonreal root.(By Jonathan)
See Also
2025 USAMO (Problems • Resources) | ||
Preceded by Problem 1 |
Followed by Problem 3 | |
1 • 2 • 3 • 4 • 5 • 6 | ||
All USAMO Problems and Solutions |
See Also
2025 USAMO (Problems • Resources) | ||
Preceded by Problem 1 |
Followed by Problem 3 | |
1 • 2 • 3 • 4 • 5 • 6 | ||
All USAMO Problems and Solutions |