Difference between revisions of "1968 IMO Problems/Problem 1"
m (→Solution 2) |
m |
||
(20 intermediate revisions by 3 users not shown) | |||
Line 20: | Line 20: | ||
+ | ==Solution 2== | ||
+ | |||
+ | (Note: this proof is an expansion by pf02 of an outline of a solution | ||
+ | posted here before.) | ||
+ | |||
+ | In a given triangle <math>ABC</math>, let <math>A=2B</math>, <math>\implies C=180-3B</math>, and <math>\sin C=\sin 3B</math>. | ||
+ | Then | ||
+ | |||
+ | <math>\sin ^2 A = \sin ^2 2B = 2 \sin B \cos B \sin 2B = \sin B(\sin B + \sin 3B) = \sin B(\sin B + \sin C)</math> | ||
+ | |||
+ | Hence, | ||
+ | |||
+ | <math>a^2 = b(b + c)\ (*)</math> | ||
+ | |||
+ | Indeed, we know from the [[Law of Sines]] that | ||
+ | |||
+ | <math>\frac{\sin A}{a} = \frac{\sin B}{b} = \frac{\sin C}{c}</math>. | ||
+ | |||
+ | Denote this ratio by <math>r</math>; we have <math>\sin A = ra, \sin B = rb, \sin C = rc</math>. | ||
+ | Substitute in <math>\sin ^2 A = \sin B(\sin B + \sin C)</math> and simplify by <math>r^2</math>. | ||
+ | We get <math>(*)</math>. | ||
+ | |||
+ | At this point, notice that <math>(*)</math> is equivalent to the equality | ||
+ | <math>a^2c = b(a^2 + c^2 - b^2)</math> from Solution 1. Indeed, the latter | ||
+ | can be rewritten as <math>a^2(c - b) = b(c + b)(c - b)</math>, and we know | ||
+ | that <math>c \ne b</math>. So we could simply quote the fact (proven in | ||
+ | Solution 1) that if <math>a, b, c</math> are consecutive integers and | ||
+ | <math>a^2 = b(c + b)</math>, then <math>b = 4, c = 5, a = 6</math> is the only solution | ||
+ | which could be the sides of a triangle. | ||
+ | |||
+ | For the sake of completeness, and for fun, I give a slightly | ||
+ | different proof here. | ||
+ | |||
+ | We have six possibilities, depending on how the three consecutive | ||
+ | numbers are ordered. The six possibilities are: | ||
+ | |||
+ | 1: <math>\ \ a = b - 2, c = b - 1, b</math> | ||
+ | |||
+ | 2: <math>\ \ c = b - 2, a = b - 1, b</math> | ||
+ | |||
+ | 3: <math>\ \ a = b - 1, b, c = b + 1</math> | ||
+ | |||
+ | 4: <math>\ \ c = b - 1, b, a = b + 1</math> | ||
+ | |||
+ | 5: <math>\ \ b, a = b + 1, c = b + 2</math> | ||
+ | |||
+ | 6: <math>\ \ b, c = b + 1, a = b + 2</math> | ||
+ | |||
+ | For each case, we could substitute <math>a, c</math> in <math>(*)</math>, get an | ||
+ | equation in <math>b</math>, solve it, and get all the possible solutions. | ||
+ | As a shortcut, notice that <math>(*)</math> implies that <math>b|a^2</math>. If | ||
+ | <math>a, b</math> are consecutive integers, then they are relatively | ||
+ | prime, so <math>b|a^2</math> can not be true unless <math>b = 1</math>. In this | ||
+ | case the triangle would have sides <math>1, 2, 3</math>, which is | ||
+ | impossible. This eliminates cases 2, 3, 4 and 5. | ||
+ | |||
+ | In case 1, <math>(*)</math> becomes | ||
+ | |||
+ | <math>(b - 2)^2 = b(b + b - 1)</math>, or <math>b^2 + 3b - 4 = 0</math>. | ||
+ | |||
+ | This has solutions <math>1, -4</math>. The value <math>b = -4</math> is impossible. | ||
+ | The value <math>b = 1</math> yields <math>a = -1, c = 0</math>, which is impossible. | ||
− | + | In case 6, <math>(*)</math> becomes | |
+ | |||
+ | <math>(b + 2)^2 = b(b + b + 1)</math>, or <math>b^2 - 3b - 4 = 0</math>. | ||
+ | |||
+ | The solutions are <math>-1, 4</math>. The value <math>b = -1</math> is impossible. | ||
+ | Thus, we get the unique triangle <math>a = 6, b = 4, c = 5</math>. | ||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
==Solution 3== | ==Solution 3== | ||
NO TRIGONOMETRY!!! | NO TRIGONOMETRY!!! | ||
− | Let <math>a, b, c</math> be the side lengths of a triangle in which <math><C = 2< | + | Let <math>a, b, c</math> be the side lengths of a triangle in which <math>\angle C = 2\angle B.</math> |
+ | |||
+ | Extend <math>AC</math> to <math>D</math> such that <math>CD = BC = a.</math> Then <math>\angle CDB = \frac{\angle ACB}{2} = \angle ABC</math>, so <math>ABC</math> and <math>ADB</math> are similar by AA Similarity. Hence, <math>c^2 = b(a+b)</math>. Then proceed as in Solution 2, as only algebraic manipulations are left. | ||
+ | |||
+ | |||
+ | ==Solution 4== | ||
+ | |||
+ | Note: Adding this 4th solution is justified by the fact | ||
+ | that it is extremely straightforward, and by the fact | ||
+ | that it shows that there are exactly two triangles for | ||
+ | which the sides differ by <math>1</math> (i.e. they are | ||
+ | <math>x, x + 1, x + 2</math> for some <math>x</math>), and the condition | ||
+ | on the angles is satisfied (that one is twice the other). | ||
+ | But only one of the solutions has <math>x</math> integer. | ||
+ | |||
+ | So, let us start by assuming that two angles are | ||
+ | <math>\alpha, 2\alpha</math> and the sides are <math>x, x + 1, x + 2</math>. | ||
+ | We will want to apply the [[Law of Sines]]: | ||
+ | |||
+ | <math>\frac{x}{\sin A} = \frac{x + 1}{\sin B} = \frac{x + 2}{\sin C}</math> | ||
+ | |||
+ | The angles <math>A, B, C</math> should be so that <math>\sin A \le \sin B \le \sin C</math>, | ||
+ | but we don't know how to map <math>\{A, B, C\}</math> to | ||
+ | <math>\{\alpha, 2\alpha, \pi - 3\alpha\}</math>. One thing we know, is that | ||
+ | <math>\sin \alpha < \sin 2\alpha</math>. Indeed, if <math>\alpha \le \pi/4</math> the | ||
+ | inequality is true because <math>\sin</math> is increasing on <math>[0, \pi/2]</math>. | ||
+ | Now note that <math>\alpha < \pi/3</math> since otherwise <math>\pi - 3\alpha</math> | ||
+ | could not be the angle of a triangle. So, if | ||
+ | <math>\pi/4 < \alpha < \pi/3</math> then <math>\pi/2 < 2\alpha < 2\pi/3</math> | ||
+ | and <math>\sin \alpha < \sqrt{3}/2 < \sin 2\alpha</math>. | ||
+ | |||
+ | That means we will have to consider three possibilities: | ||
+ | |||
+ | 1: <math>\ \ \frac{x}{\sin \alpha} = \frac{x + 1}{\sin 2\alpha} = \frac{x + 2}{\sin (\pi - 3\alpha)}</math> | ||
+ | |||
+ | 2: <math>\ \ \frac{x}{\sin \alpha} = \frac{x + 1}{\sin (\pi - 3\alpha)} = \frac{x + 2}{\sin 2\alpha}</math> | ||
+ | |||
+ | 3: <math>\ \ \frac{x}{\sin (\pi - 3\alpha)} = \frac{x + 1}{\sin \alpha} = \frac{x + 2}{\sin 2\alpha}</math> | ||
+ | |||
+ | Using the identities | ||
+ | <math>\sin (\pi - \theta) = \sin \theta, \sin 2\theta = 2\sin \theta \cos \theta</math> | ||
+ | and | ||
+ | <math>\sin 3\theta = 3\sin \theta - 4\sin^3 \theta = \sin \theta\ (4\cos^2 \theta - 1)</math> | ||
+ | and simplifying by <math>\sin \alpha</math> the three cases become | ||
+ | |||
+ | 1: <math>\ \ x = \frac{x + 1}{2\cos \alpha} = \frac{x + 2}{4\cos^2 \alpha - 1}</math> | ||
+ | |||
+ | 2: <math>\ \ x = \frac{x + 1}{4\cos^2 \alpha - 1} = \frac{x + 2}{2\cos \alpha}</math> | ||
+ | |||
+ | 3: <math>\ \ \frac{x}{4\cos^2 \alpha - 1} = x + 1 = \frac{x + 2}{2\cos \alpha}</math> | ||
+ | |||
+ | Each case is a system of two equations in two unknowns, <math>x, \cos \alpha</math>. | ||
+ | We will solve each system, obtain all possible solutions, and chose those | ||
+ | values for which <math>x, x + 1, x + 2</math> and <math>\alpha, 2\alpha, \pi - 3\alpha</math> | ||
+ | can be the sides and angles of a triangle. | ||
+ | |||
+ | Case 1: Compute <math>x</math> from <math>x = \frac{x + 1}{2\cos \alpha}</math>. We get | ||
+ | <math>x = \frac{1}{2\cos \alpha - 1}</math>. Substitute <math>x</math> in | ||
+ | <math>x = \frac{x + 2}{4\cos^2 \alpha - 1}</math>. After doing all the computations | ||
+ | we get <math>4\cos^2 \alpha - 4\cos \alpha = 0</math>. The roots are <math>\cos \alpha = 0</math> | ||
+ | and <math>\cos \alpha = 1</math>. None are acceptable if <math>\alpha</math> is an angle of a | ||
+ | triangle. So case 1 yields no solutions. | ||
+ | |||
+ | Case 2: Compute <math>x</math> from <math>x = \frac{x + 2}{2\cos \alpha}</math>. We get | ||
+ | <math>x = \frac{2}{2\cos \alpha - 1}</math>. Substitute <math>x</math> in | ||
+ | <math>x = \frac{x + 1}{4\cos^2 \alpha - 1}</math>. After doing all the computations | ||
+ | we get <math>8\cos^2 \alpha - 2\cos \alpha - 3 = 0</math>. The solutions are | ||
+ | <math>\cos \alpha = \frac{3}{4}</math> and <math>\cos \alpha = -\frac{1}{2}</math>. Only | ||
+ | <math>\cos \alpha = \frac{3}{4}</math> is acceptable, and it yields | ||
+ | <math>x = \frac{2}{\frac{6}{4} - 1} = 4</math>. Thus <math>4, 5, 6</math> is a possible | ||
+ | solution to the problem. | ||
+ | |||
+ | Case 3: Compute <math>x</math> from <math>x + 1 = \frac{x + 2}{2\cos \alpha}</math>. We get | ||
+ | <math>x = \frac{2 - 2\cos \alpha}{2\cos \alpha - 1}</math>. Substitute <math>x</math> in | ||
+ | <math>x + 1 = \frac{x}{4\cos^2 \alpha - 1}</math>. After doing all the computations | ||
+ | we get <math>4\cos^2 \alpha + 2\cos \alpha - 3 = 0</math>. The roots are | ||
+ | <math>\cos \alpha = \frac{-1 \pm \sqrt{13}}{4} \approx 0.65, -1.15</math>. | ||
+ | The positive value is for an <math>\alpha</math> acceptable as an angle in | ||
+ | a triangle (it is a little over <math>\pi/4</math>), and yields | ||
+ | <math>x = \frac{\sqrt{13} + 1}{2}</math>. We can easily verify that | ||
+ | <math>x, x + 1, x + 2</math> can be the sides of a triangle. (Indeed, | ||
+ | they are | ||
+ | <math>\frac{\sqrt{13} + 1}{2}, \frac{\sqrt{13} + 3}{2}, \frac{\sqrt{13} + 5}{2}</math> | ||
+ | and | ||
+ | <math>\frac{\sqrt{13} + 1}{2} + \frac{\sqrt{13} + 3}{2} > \frac{\sqrt{13} + 5}{2}</math>). | ||
+ | However, they are not integer, so are not solutions to the problem. | ||
+ | |||
+ | The only solution to the problem is the triangle with sides <math>4, 5, 6</math> | ||
+ | from case 2. | ||
+ | |||
+ | [Solution by pf02, August 2024] | ||
− | |||
==See Also== | ==See Also== | ||
{{IMO box|year=1968|before=First Problem|num-a=2}} | {{IMO box|year=1968|before=First Problem|num-a=2}} | ||
− | + | ||
[[Category:Olympiad Geometry Problems]] | [[Category:Olympiad Geometry Problems]] |
Latest revision as of 18:24, 10 November 2024
Problem
Prove that there is one and only one triangle whose side lengths are consecutive integers, and one of whose angles is twice as large as another.
Solution 1
In triangle , let
,
,
,
, and
. Using the Law of Sines gives that
Therefore . Using the Law of Cosines gives that
This can be simplified to . Since
,
, and
are positive integers,
. Note that if
is between
and
, then
is relatively prime to
and
, and
cannot possibly divide
. Therefore
is either the least of the three consecutive integers or the greatest.
Assume that is the least of the three consecutive integers. Then either
or
, depending on if
or
. If
, then
is 1 or 2.
couldn't be 1, for if it was then the triangle would be degenerate. If
is 2, then
, but
and
must be 3 and 4 in some order, which means that this triangle doesn't exist. therefore
cannot divide
, and so
must divide
. If
then
, so
is 1, 2, or 4. Clearly
cannot be 1 or 2, so
must be 4. Therefore
. This shows that
and
, and the triangle has sides that measure 4, 5, and 6.
Now assume that is the greatest of the three consecutive integers. Then either
or
, depending on if
or
.
is absurd, so
, and
. Therefore
is 1, 2, or 4. However, all of these cases are either degenerate or have been previously ruled out, so
cannot be the greatest of the three consecutive integers.
This shows that there is exactly one triangle with this property - and it has side lengths of 4, 5, and 6.
Solution 2
(Note: this proof is an expansion by pf02 of an outline of a solution posted here before.)
In a given triangle , let
,
, and
.
Then
Hence,
Indeed, we know from the Law of Sines that
.
Denote this ratio by ; we have
.
Substitute in
and simplify by
.
We get
.
At this point, notice that is equivalent to the equality
from Solution 1. Indeed, the latter
can be rewritten as
, and we know
that
. So we could simply quote the fact (proven in
Solution 1) that if
are consecutive integers and
, then
is the only solution
which could be the sides of a triangle.
For the sake of completeness, and for fun, I give a slightly different proof here.
We have six possibilities, depending on how the three consecutive numbers are ordered. The six possibilities are:
1:
2:
3:
4:
5:
6:
For each case, we could substitute in
, get an
equation in
, solve it, and get all the possible solutions.
As a shortcut, notice that
implies that
. If
are consecutive integers, then they are relatively
prime, so
can not be true unless
. In this
case the triangle would have sides
, which is
impossible. This eliminates cases 2, 3, 4 and 5.
In case 1, becomes
, or
.
This has solutions . The value
is impossible.
The value
yields
, which is impossible.
In case 6, becomes
, or
.
The solutions are . The value
is impossible.
Thus, we get the unique triangle
.
Solution 3
NO TRIGONOMETRY!!!
Let be the side lengths of a triangle in which
Extend to
such that
Then
, so
and
are similar by AA Similarity. Hence,
. Then proceed as in Solution 2, as only algebraic manipulations are left.
Solution 4
Note: Adding this 4th solution is justified by the fact
that it is extremely straightforward, and by the fact
that it shows that there are exactly two triangles for
which the sides differ by (i.e. they are
for some
), and the condition
on the angles is satisfied (that one is twice the other).
But only one of the solutions has
integer.
So, let us start by assuming that two angles are
and the sides are
.
We will want to apply the Law of Sines:
The angles should be so that
,
but we don't know how to map
to
. One thing we know, is that
. Indeed, if
the
inequality is true because
is increasing on
.
Now note that
since otherwise
could not be the angle of a triangle. So, if
then
and
.
That means we will have to consider three possibilities:
1:
2:
3:
Using the identities
and
and simplifying by
the three cases become
1:
2:
3:
Each case is a system of two equations in two unknowns, .
We will solve each system, obtain all possible solutions, and chose those
values for which
and
can be the sides and angles of a triangle.
Case 1: Compute from
. We get
. Substitute
in
. After doing all the computations
we get
. The roots are
and
. None are acceptable if
is an angle of a
triangle. So case 1 yields no solutions.
Case 2: Compute from
. We get
. Substitute
in
. After doing all the computations
we get
. The solutions are
and
. Only
is acceptable, and it yields
. Thus
is a possible
solution to the problem.
Case 3: Compute from
. We get
. Substitute
in
. After doing all the computations
we get
. The roots are
.
The positive value is for an
acceptable as an angle in
a triangle (it is a little over
), and yields
. We can easily verify that
can be the sides of a triangle. (Indeed,
they are
and
).
However, they are not integer, so are not solutions to the problem.
The only solution to the problem is the triangle with sides
from case 2.
[Solution by pf02, August 2024]
See Also
1968 IMO (Problems) • Resources | ||
Preceded by First Problem |
1 • 2 • 3 • 4 • 5 • 6 | Followed by Problem 2 |
All IMO Problems and Solutions |