Difference between revisions of "1961 IMO Problems/Problem 3"
m |
|||
(One intermediate revision by the same user not shown) | |||
Line 7: | Line 7: | ||
where <math>n</math> is a given positive integer. | where <math>n</math> is a given positive integer. | ||
+ | ==Solution 1== | ||
+ | Since <math>cos^2x + sin^2x = 1</math>, we cannot have solutions with <math>n\ne2</math> and <math>0<|cos(x)|,|sin(x)|<1</math>. Nor can we have solutions with <math>n=2</math>, because the sign is wrong. So the only solutions have <math>sin (x) = 0</math> or <math>cos (x) = 0</math>, and these are: <math>x =</math> multiple of <math>\pi</math>, and <math>n</math> even; <math>x </math> even multiple of <math>\pi</math> and <math>n</math> odd; <math>x</math> = even multiple of <math>\pi + \frac{3\pi}{2}</math> and <math>n</math> odd. | ||
− | ==Solution== | + | ==Solution 2 (Calculus)== |
+ | First consider <math>n=1</math>. Squaring both sides yields <math>\cos^2{x}-2\cos{x}\sin{x}+\sin^2{x}=1</math>, and utilizing identities yields <math>\sin{2x}=0</math>. This results in solutions <math>x=\frac{\pi}{2}k</math> for integers <math>k</math>. After manually checking the solutions in the original equation, we conclude that only <math>x=2k\pi</math> and <math>x=\frac{3\pi}{2}+2k\pi</math> are solutions to the equation. | ||
− | |||
+ | Next, we consider <math>n=2</math>. This is equivalent to <math>\cos{2x}=1</math>, so the solutions are <math>x=k\pi</math> for integers <math>k</math>. Once again, substituting into the equation results in a valid identity, so these are the only solutions. | ||
+ | |||
+ | |||
+ | Let <math>f(x)=\cos^n{x}-\sin^n{x}</math>. Taking the derivative yields <math>f'(x)=-n\sin{x}\cos{x}(\sin^{n-2}{x}+\cos^{n-2}{x})</math> for <math>n>2</math>. We wish to consider extrema, so we find all <math>x</math>-values such that <math>f'(x)=0</math> by checking the zeros of each factor. In <math>x\in[0,2\pi]</math>, <math>f'(x)=0</math> for even <math>n</math> when <math>x=0,\frac{\pi}{2},\pi,\frac{3\pi}{2},2\pi</math>, and for odd <math>n</math> when <math>x=0,\frac{\pi}{2},\frac{3\pi}{4},\pi,\frac{3\pi}{2},\frac{7\pi}{4},2\pi</math>. The differences between the two come from the <math>\sin^{n-2}{x}+\cos^{n-2}{x}</math> factor, reaching zeros when <math>\tan^{n-2}{x}=-1</math>. This is impossible when <math>n</math> is even, and if <math>n</math> is odd, it implies that <math>\tan{x}=-1</math>, hence the solutions. | ||
+ | |||
+ | |||
+ | We then calculate the value of <math>f(x)</math> at every such <math>x</math>. For even <math>n</math>, the <math>x</math>-values <math>0,\frac{\pi}{2},\pi,\frac{3\pi}{2},2\pi</math> result in <math>f(x)</math>-values of <math>1,-1,1,-1,1</math>. Since these are the only possible extrema, we conclude that <math>1</math> is always a maximum and <math>-1</math> is always a minimum; thus the only solutions are the maxima, which occur at <math>x=0,\pi,2\pi</math> when <math>x\in[0,2\pi]</math>; more generally, we have <math>x=k\pi</math> for integers <math>k</math>. | ||
+ | |||
+ | |||
+ | Next, for odd <math>n</math>, the <math>x</math>-values <math>x=0,\frac{\pi}{2},\frac{3\pi}{4},\pi,\frac{3\pi}{2},\frac{7\pi}{4},2\pi</math> yield <math>f(x)</math>-values of <math>1,-1,-2\left(\frac{1}{\sqrt{2}}\right)^n,-1,1,2\left(\frac{1}{\sqrt{2}}\right)^n,1</math>. Obviously <math>x=0,\frac{3\pi}{2},2\pi</math> are all solutions; we show that on our closed domain these are the only solutions. On <math>x=0</math> to <math>x=\frac{\pi}{2}</math>, the function is always decreasing; if it were increasing at any point, an extremum would be required to change the sign of the derivative; however, there does not exist an extremum between these two points (since we have already found all the extrema!); thus only <math>x=0</math> satisfies the equation. We can repeat this argument for all values between two extrema; eventually, we can conclude that the only times that <math>f(x)=1</math> are the above (since no segment contains <math>1</math> at any point other than an endpoint; as a result, the function cannot equal <math>1</math> at any point other than an endpoint of some segment, which are the extrema). As a result, extending this to all <math>x</math>, the solutions are <math>x=2k\pi</math> and <math>x=\frac{3\pi}{2}+2k\pi</math> for integers <math>k</math>. | ||
+ | |||
+ | |||
+ | Since these results match our <math>n=1</math> and <math>n=2</math> cases, we conclude that when <math>n</math> is odd, the solutions are <math>x=2k\pi</math> and <math>x=\frac{3\pi}{2}+2k\pi</math> for integers <math>k</math>, and when <math>n</math> is even, the solutions are <math>x=k\pi</math> for integers <math>k</math>. | ||
+ | |||
+ | ~ [https://artofproblemsolving.com/wiki/index.php/User:Eevee9406 eevee9406] | ||
{{IMO box|year=1961|num-b=2|num-a=4}} | {{IMO box|year=1961|num-b=2|num-a=4}} |
Latest revision as of 23:07, 10 July 2025
Problem
Solve the equation
where is a given positive integer.
Solution 1
Since , we cannot have solutions with
and
. Nor can we have solutions with
, because the sign is wrong. So the only solutions have
or
, and these are:
multiple of
, and
even;
even multiple of
and
odd;
= even multiple of
and
odd.
Solution 2 (Calculus)
First consider . Squaring both sides yields
, and utilizing identities yields
. This results in solutions
for integers
. After manually checking the solutions in the original equation, we conclude that only
and
are solutions to the equation.
Next, we consider . This is equivalent to
, so the solutions are
for integers
. Once again, substituting into the equation results in a valid identity, so these are the only solutions.
Let . Taking the derivative yields
for
. We wish to consider extrema, so we find all
-values such that
by checking the zeros of each factor. In
,
for even
when
, and for odd
when
. The differences between the two come from the
factor, reaching zeros when
. This is impossible when
is even, and if
is odd, it implies that
, hence the solutions.
We then calculate the value of at every such
. For even
, the
-values
result in
-values of
. Since these are the only possible extrema, we conclude that
is always a maximum and
is always a minimum; thus the only solutions are the maxima, which occur at
when
; more generally, we have
for integers
.
Next, for odd , the
-values
yield
-values of
. Obviously
are all solutions; we show that on our closed domain these are the only solutions. On
to
, the function is always decreasing; if it were increasing at any point, an extremum would be required to change the sign of the derivative; however, there does not exist an extremum between these two points (since we have already found all the extrema!); thus only
satisfies the equation. We can repeat this argument for all values between two extrema; eventually, we can conclude that the only times that
are the above (since no segment contains
at any point other than an endpoint; as a result, the function cannot equal
at any point other than an endpoint of some segment, which are the extrema). As a result, extending this to all
, the solutions are
and
for integers
.
Since these results match our and
cases, we conclude that when
is odd, the solutions are
and
for integers
, and when
is even, the solutions are
for integers
.
1961 IMO (Problems) • Resources | ||
Preceded by Problem 2 |
1 • 2 • 3 • 4 • 5 • 6 | Followed by Problem 4 |
All IMO Problems and Solutions |