Difference between revisions of "1990 USAMO Problems/Problem 2"
(solution 2) |
(solution 2) |
||
Line 30: | Line 30: | ||
== Solution 2 == | == Solution 2 == | ||
− | We claim that the only solution is <math>\boxed{x=4}</math> for all such <math>n</math>. To show this, we consider all <math>x</math> and find solutions <math>n</math>. | + | We claim that the only solution is <math>\boxed{x=4}</math> for all such <math>n</math>. To show this, we consider all <math>x</math> and find solutions <math>n</math>. |
Before we consider solutions, we show that for all <math>x</math>, <math>f_n(x)</math> is positive for all positive integers <math>n</math> by induction. For our base case: | Before we consider solutions, we show that for all <math>x</math>, <math>f_n(x)</math> is positive for all positive integers <math>n</math> by induction. For our base case: | ||
Line 36: | Line 36: | ||
so it is positive. Next, for our inductive step, assume for some <math>n=k-1</math> that <math>f_{k-1}(x)</math> is positive; thus <math>f_{k-1}(x)>0</math>, and we show that <math>f_k(x)</math> is positive: | so it is positive. Next, for our inductive step, assume for some <math>n=k-1</math> that <math>f_{k-1}(x)</math> is positive; thus <math>f_{k-1}(x)>0</math>, and we show that <math>f_k(x)</math> is positive: | ||
<cmath>f_k(x)=\sqrt{x^2+6f_{k-1}(x)}\ge\sqrt{0+6f_{k-1}(x)}>\sqrt{0+0}=0</cmath> | <cmath>f_k(x)=\sqrt{x^2+6f_{k-1}(x)}\ge\sqrt{0+6f_{k-1}(x)}>\sqrt{0+0}=0</cmath> | ||
− | thus <math>f_k(x)>0</math>, so we have proven the claim by induction. | + | thus <math>f_k(x)>0</math>, so we have proven the claim by induction. |
− | First, we consider negative <math>x</math>. We know that for negative <math>x</math>, if the equation were to have a solution, we would have <math>f_n(x)=2x<0</math> for some <math>n</math>. However, <math>f_n(x)</math> is always nonnegative since <math>f_n(x)</math> is always the square root of some number, which can never be negative; thus there are no solutions in this case. | + | First, we consider negative <math>x</math>. We know that for negative <math>x</math>, if the equation were to have a solution, we would have <math>f_n(x)=2x<0</math> for some <math>n</math>. However, <math>f_n(x)</math> is always nonnegative since <math>f_n(x)</math> is always the square root of some number, which can never be negative; thus there are no solutions in this case. |
− | Next, we consider <math>x=0</math>. Solutions would have to satisfy <math>f_n(0)=0</math>; we have previously shown that all <math>f_n(x)</math> are positive, so there are no such <math>n</math>. | + | Next, we consider <math>x=0</math>. Solutions would have to satisfy <math>f_n(0)=0</math>; we have previously shown that all <math>f_n(x)</math> are positive, so there are no such <math>n</math>. |
Finally, we consider positive <math>x</math>. We divide this set into three groups: <math>x<4</math>, <math>x=4</math>, and <math>x>4</math>. We claim that for increasing <math>n</math>, <math>f_n(x)</math> is decreasing, constant, and increasing, respectively, and we prove this using induction. First, we analyze <math>x<4</math>. Then | Finally, we consider positive <math>x</math>. We divide this set into three groups: <math>x<4</math>, <math>x=4</math>, and <math>x>4</math>. We claim that for increasing <math>n</math>, <math>f_n(x)</math> is decreasing, constant, and increasing, respectively, and we prove this using induction. First, we analyze <math>x<4</math>. Then | ||
Line 48: | Line 48: | ||
Thus <math>f_2(x)<f_1(x)</math>. For our inductive step, if for some <math>n=k-1</math> we have <math>f_n(x)<f_{n-1}(x)</math>, we show that <math>f_{n+1}(x)<f_n(x)</math>: | Thus <math>f_2(x)<f_1(x)</math>. For our inductive step, if for some <math>n=k-1</math> we have <math>f_n(x)<f_{n-1}(x)</math>, we show that <math>f_{n+1}(x)<f_n(x)</math>: | ||
<cmath>f_{n+1}(x)=\sqrt{x^2+6f_n(x)}<\sqrt{x^2+6f_{n-1}(x)}=f_n(x)</cmath> | <cmath>f_{n+1}(x)=\sqrt{x^2+6f_n(x)}<\sqrt{x^2+6f_{n-1}(x)}=f_n(x)</cmath> | ||
− | Thus <math>f_1(x)>f_2(x)>f_3(x)>\ldots</math> and the conclusion follows. | + | Thus <math>f_1(x)>f_2(x)>f_3(x)>\ldots</math> and the conclusion follows. |
Next, for <math>x=4</math>, <math>f_1(4)=8</math> by substituting. Then, if <math>f_{n-1}(4)=8</math>, we show that <math>f_n(4)=8</math> as well: | Next, for <math>x=4</math>, <math>f_1(4)=8</math> by substituting. Then, if <math>f_{n-1}(4)=8</math>, we show that <math>f_n(4)=8</math> as well: | ||
<math>f_n(4)=\sqrt{x^2+6f_{n-1}(x)}=\sqrt{16+6\cdot8}=\sqrt{64}=8</math> | <math>f_n(4)=\sqrt{x^2+6f_{n-1}(x)}=\sqrt{16+6\cdot8}=\sqrt{64}=8</math> | ||
− | so the sequence is constant. Notice that in this case, the equation we must solve becomes <math>f_n(4)=8</math>, which is true for all positive integers <math>n</math>. | + | so the sequence is constant. Notice that in this case, the equation we must solve becomes <math>f_n(4)=8</math>, which is true for all positive integers <math>n</math>. |
Finally, we analyze <math>x>4</math>. Then | Finally, we analyze <math>x>4</math>. Then | ||
Line 60: | Line 60: | ||
Thus <math>f_2(x)>f_1(x)</math>. For our inductive step, if for some <math>n=k-1</math> we have <math>f_n(x)>f_{n-1}(x)</math>, we show that <math>f_{n+1}(x)>f_n(x)</math>: | Thus <math>f_2(x)>f_1(x)</math>. For our inductive step, if for some <math>n=k-1</math> we have <math>f_n(x)>f_{n-1}(x)</math>, we show that <math>f_{n+1}(x)>f_n(x)</math>: | ||
<cmath>f_{n+1}(x)=\sqrt{x^2+6f_n(x)}>\sqrt{x^2+6f_{n-1}(x)}=f_n(x)</cmath> | <cmath>f_{n+1}(x)=\sqrt{x^2+6f_n(x)}>\sqrt{x^2+6f_{n-1}(x)}=f_n(x)</cmath> | ||
− | Thus <math>f_1(x)<f_2(x)<f_3(x)<\ldots</math> and the conclusion follows. | + | Thus <math>f_1(x)<f_2(x)<f_3(x)<\ldots</math> and the conclusion follows. |
If for some positive integer <math>n</math> we have <math>f_n(x)=2x</math>, then: | If for some positive integer <math>n</math> we have <math>f_n(x)=2x</math>, then: | ||
+ | |||
\begin{align*} | \begin{align*} | ||
− | + | f_n(x)&=\sqrt{x^2+6f_{n-1}(x)}\\ | |
− | + | 2x&=\sqrt{x^2+6f_{n-1}(x)}\\ | |
− | + | 4x^2&=x^2+6f_{n-1}(x)\\ | |
− | + | 3x^2&=6f_{n-1}(x)\\ | |
− | + | \frac{1}{2}x^2&=f_{n-1}(x) | |
\end{align*} | \end{align*} | ||
− | However, if <math>0<x<4</math> is a solution, then we must have <math>f_{n-1}(x)>f_n(x)</math>; substituting values yields <math>\frac{1}{2}x^2>2x</math>, which implies <math>x>4</math>; this is a contradiction. Similarly, if <math>x>4</math> is a solution, then we must have <math>f_{n-1}(x)<f_n(x)</math>; substituting values yields <math>\frac{1}{2}x^2<2x</math>, which implies <math>x<4</math>; this is also a contradiction. | + | However, if <math>0<x<4</math> is a solution, then we must have <math>f_{n-1}(x)>f_n(x)</math>; substituting values yields <math>\frac{1}{2}x^2>2x</math>, which implies <math>x>4</math>; this is a contradiction. Similarly, if <math>x>4</math> is a solution, then we must have <math>f_{n-1}(x)<f_n(x)</math>; substituting values yields <math>\frac{1}{2}x^2<2x</math>, which implies <math>x<4</math>; this is also a contradiction. |
As a result, since we have already established that <math>x=4</math> is a solution for all <math>n</math> and no other <math>x</math> can be solutions, we conclude that for each positive integer <math>n</math>, the only real solution of the equation is <math>x=4</math>. | As a result, since we have already established that <math>x=4</math> is a solution for all <math>n</math> and no other <math>x</math> can be solutions, we conclude that for each positive integer <math>n</math>, the only real solution of the equation is <math>x=4</math>. |
Revision as of 15:45, 16 March 2025
Contents
Problem
A sequence of functions is defined recursively as follows:
(Recall that
is understood to represent the positive square root.) For each positive integer
, find all real solutions of the equation
.
Solution 1
We define . Then the recursive relation holds for
, as well.
Since for all nonnegative integers
, it suffices to consider nonnegative values of
.
We claim that the following set of relations hold true for all natural numbers and nonnegative reals
:
To prove this claim, we induct on
. The statement evidently holds for our base case,
.
Now, suppose the claim holds for . Then
The claim therefore holds by induction. It then follows that for all nonnegative integers
,
is the unique solution to the equation
.
Solution 2
We claim that the only solution is for all such
. To show this, we consider all
and find solutions
.
Before we consider solutions, we show that for all ,
is positive for all positive integers
by induction. For our base case:
so it is positive. Next, for our inductive step, assume for some
that
is positive; thus
, and we show that
is positive:
thus
, so we have proven the claim by induction.
First, we consider negative . We know that for negative
, if the equation were to have a solution, we would have
for some
. However,
is always nonnegative since
is always the square root of some number, which can never be negative; thus there are no solutions in this case.
Next, we consider . Solutions would have to satisfy
; we have previously shown that all
are positive, so there are no such
.
Finally, we consider positive . We divide this set into three groups:
,
, and
. We claim that for increasing
,
is decreasing, constant, and increasing, respectively, and we prove this using induction. First, we analyze
. Then
For our base case:
Thus
. For our inductive step, if for some
we have
, we show that
:
Thus
and the conclusion follows.
Next, for ,
by substituting. Then, if
, we show that
as well:
so the sequence is constant. Notice that in this case, the equation we must solve becomes
, which is true for all positive integers
.
Finally, we analyze . Then
For our base case:
Thus
. For our inductive step, if for some
we have
, we show that
:
Thus
and the conclusion follows.
If for some positive integer we have
, then:
\begin{align*}
f_n(x)&=\sqrt{x^2+6f_{n-1}(x)}\\
2x&=\sqrt{x^2+6f_{n-1}(x)}\\
4x^2&=x^2+6f_{n-1}(x)\\
3x^2&=6f_{n-1}(x)\\
\frac{1}{2}x^2&=f_{n-1}(x)
\end{align*}
However, if is a solution, then we must have
; substituting values yields
, which implies
; this is a contradiction. Similarly, if
is a solution, then we must have
; substituting values yields
, which implies
; this is also a contradiction.
As a result, since we have already established that is a solution for all
and no other
can be solutions, we conclude that for each positive integer
, the only real solution of the equation is
.
See Also
1990 USAMO (Problems • Resources) | ||
Preceded by Problem 1 |
Followed by Problem 3 | |
1 • 2 • 3 • 4 • 5 | ||
All USAMO Problems and Solutions |
These problems are copyrighted © by the Mathematical Association of America, as part of the American Mathematics Competitions.