Individual Integral Triangle

Geometry Level 4

Given a circumradius 9 9 and the inradius 4 4 , there exist some triangles with an integer area A A . Find A A .


The answer is 90.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

1 solution

Chris Lewis
Apr 14, 2021

Notation: R R is the circumradius; r r the inradius); O O the circumcentre; I I the incentre; A , B , C A,B,C are the vertices of the triangle (and also the angles at those vertices); a , b , c a,b,c are the lengths of the sides opposite those vertices; T T is the area of the triangle.

By Euler's theorem on triangles, the circumcentre and the incentre are a distance d = R 2 2 R r d=\sqrt{R^2-2Rr} apart.

Take O O as an origin; let A A have coordinates ( R , 0 ) (R,0) and I ( d cos ( θ ) , d sin ( θ ) ) I(d \cos(\theta),d \sin(\theta)) . Then all the different triangles satisfying the conditions are found by varying θ \theta .

Since the incentre lies on the angle bisectors, I A B = A 2 \angle IAB=\frac{A}{2} ; so sin A 2 = r A I \sin \frac{A}{2} = \frac{r}{AI}

By the extended sine rule, a = 2 R sin A a=2R \sin A .

Now consider the triangle B I C BIC . By definition of the incentre, the altitude from I I has length r r . Also it's easy to show that B I C = A + π 2 \angle BIC = \frac{A+\pi}{2} . Again using the fact the incentre lies on the angle bisectors, I B C = B 2 \angle IBC=\frac{B}{2} .

Combining all of this, B C = a = 2 R sin A = r ( cot B 2 + cot C 2 ) = r ( cot B 2 + tan A + B 2 ) BC=a=2R\sin A=r\left(\cot \frac{B}{2}+\cot \frac{C}{2}\right)=r\left(\cot \frac{B}{2}+\tan \frac{A+B}{2}\right)

This can be written as a quadratic in cot B 2 \cot\frac{B}{2} .

After a LOT of trig (please comment if there's a shortcut to this), we get b c = 4 r R + A I 2 bc=4rR + AI^2

The area is given by T = 1 2 b c sin A = r ( 4 r R A I 2 + 1 ) A I 2 r 2 T=\frac12 bc \sin A = r\left(\frac{4rR}{AI^2} + 1\right) \sqrt{AI^2-r^2}

Now, we can write all this in terms of θ \theta , or note that A I 2 AI^2 varies from R d R-d to R + d R+d . Letting x = A I 2 x=AI^2 , we have T = r ( 4 r R x + 1 ) x r 2 T=r\left(\frac{4rR}{x} + 1\right) \sqrt{x-r^2}

Differentiating, d T d x = r ( 8 r 3 R 4 r R x + x 2 ) 2 x 2 x r 2 \frac{dT}{dx}=\frac{r\left(8r^3 R-4rRx+x^2 \right)}{2x^2 \sqrt{x-r^2}}

Setting the derivative equal to zero, 8 r 3 R 4 r R x + x 2 = 0 8r^3 R-4rRx+x^2=0

With R = 9 R=9 and r = 4 r=4 , this is x 2 144 x + 4608 = 0 x^2-144x+4608=0

with roots x = 48 x=48 and x = 96 x=96 . The corresponding areas are T max = 64 2 90.51 T_\text{max}=64\sqrt2\approx90.51 and T min = 40 5 89.44 T_\text{min}=40\sqrt5\approx89.44

Since there is just one integer in the interval [ T min , T max ] [T_\text{min},T_\text{max}] , this is the answer we're looking for, namely 90 \boxed{90} .


In fact, we can solve this in general; we get T min = r ( 3 R + d ) r ( 2 R r + 2 d ) R + d T_\text{min} = \frac{r (3 R + d) \sqrt{r (2R -r+ 2 d)}}{R + d} and T max = r ( 3 R d ) r ( 2 R r 2 d ) R d T_\text{max} = \frac{r (3 R - d) \sqrt{r (2R -r- 2 d)}}{R - d}

Can you find a triangle with circumradius, inradius and area of 9, 4 and 90, respectively?

Pi Han Goh - 1 month, 4 weeks ago

Log in to reply

Good point, it's not really complete without that (although I suppose continuity is obvious so we can use the intermediate value theorem).

With R = 9 R=9 , r = 4 r=4 , T = 90 T=90 , in the area formula above we get 90 = 4 ( 144 x + 1 ) x 16 90=4\left(\frac{144}{x}+1\right)\sqrt{x-16}

Squaring, multiplying through, tidying, we get (the slightly ridiculous) 4 x 3 937 x 2 + 64512 x 1327104 = 0 4 x^3-937x^2 + 64512 x -1327104 = 0

This has three real roots (as we might expect - the triangle could be rotated/reflected, changing its position without changing its size). They're not nice, though! Feel free to find the exact forms.

The smallest root is around x = 38.901 x=38.901

Recall x = A I 2 x=AI^2 ; so we have A I AI , from which we can work out the side a a ; we get a 17.7145 a\approx 17.7145

Rather than go through the whole process of finding b , c b,c via trig, we can actually just find out what sidelengths we get with the different roots of the cubic. These turn out to be around 15.4382 15.4382 and 11.8473 11.8473 ; and you can check that the triangle with these three sides works.

I'm a bit surprised to see that (up to reflection) this triangle is unique. @Pi Han Goh - you're normally very good at dealing with cubics, is there a way of writing these sidelengths exactly?

Chris Lewis - 1 month, 4 weeks ago

Log in to reply

Let z = 1 3 tan 1 ( 8 9 23 3 ) , z = \dfrac13 \tan^{-1} \left( \dfrac89 \sqrt{ \dfrac{23}3} \right) , then the side lengths of this triangle are:

  • 15 35 3 cos ( z ) 11.8473 15 - \sqrt{\dfrac{35}3} \cos (z) \approx 11.8473 ,

  • 15 + 105 6 cos ( z ) 35 2 sin ( z ) 15.438 , 15 + \dfrac{\sqrt{105}}{6} \cos(z) - \dfrac{\sqrt{35}}{2} \sin(z) \approx 15.438, and

  • 15 + 35 12 cos ( z ) + 35 2 sin ( z ) 17.714 15 + \sqrt{\dfrac{35}{12}} \cos(z) + \dfrac{\sqrt{35}}{2} \sin(z) \approx 17.714


How did I do it?

Because sin ( A 2 ) = r A I = 4 x A = 2 sin 1 ( 4 x ) \sin \left(\dfrac A2 \right) = \frac r{AI} = \dfrac 4{\sqrt x} \quad \Rightarrow \quad A = 2 \sin^{-1} \left( \dfrac 4 {\sqrt{x}} \right)

Then, a = 2 R sin ( A ) = 2 9 sin [ 2 sin 1 ( 4 x ) ] = 144 x 16 x a = 2R \sin(A) = 2\cdot9 \cdot \sin \left [2 \sin^{-1} \left( \dfrac 4 {\sqrt{x}} \right) \right ] = \dfrac{144\sqrt{x-16}}{x}

Thus, if 4 x 3 937 x 2 + 64512 x 1327104 = 0 , 4x^3 -937x^2 +64512x-1327104 = 0 , then the side lengths of this triangle are the three possible values of 144 x 16 x . \dfrac{144\sqrt{x-16}}{x}.

Pi Han Goh - 1 month, 4 weeks ago

Log in to reply

@Pi Han Goh Ugh, where's the "brilliant" button for comments when you need it? That's very clever indeed.

I'm really interested also in whether there is some obvious reason why b c = 4 r R + A I 2 bc=4rR+AI^2

I spent way too long working that out - mainly because Wolfram|Alpha told me to take my page-long equations elsewhere - but it's such a neat expression it'd be nice to explain it more intuitively. Any thoughts?

By the way, I've just added in some formulas for the min/max areas in general. It turns out these work out fairly nicely too.

Chris Lewis - 1 month, 4 weeks ago

Log in to reply

@Chris Lewis From here , R = a b c 4 r s 4 r R = a b c s = 2 a b c a + b + c . (1) R = \dfrac{abc}{4rs} \quad \Leftrightarrow \quad 4rR = \dfrac{abc}{s} = \dfrac{2abc}{a + b + c} .\tag1

And from here , A I = c sin ( B / 2 ) cos ( C / 2 ) ( A I ) 2 = c 2 2 sin 2 ( B / 2 ) 2 cos 2 ( C / 2 ) = c 2 1 cos ( B ) 1 + cos ( C ) (2) AI = c\cdot \dfrac{ \sin(B/2)}{\cos(C/2)} \quad \Leftrightarrow \quad (AI)^2 = c^2 \cdot \dfrac{ 2\sin^2(B/2)}{2\cos^2(C/2)} = c^2 \cdot \dfrac{1- \cos (B)}{1 + \cos(C)}\tag2

Using cosine rule , cos ( B ) = a 2 + c 2 b 2 2 a c , cos ( C ) = a 2 + b 2 c 2 2 a b (3) \cos(B) = \dfrac{a^2 + c^2-b^2}{2ac}, \quad\quad\quad \cos(C) = \dfrac{a^2 + b^2 - c^2}{2ab }\tag3

Substituting ( 1 ) , ( 2 ) (1), (2) and ( 3 ) (3) into 4 r R + ( A I ) 2 4rR + (AI)^2 simplifies to b c . bc.

Pi Han Goh - 1 month, 4 weeks ago

Log in to reply

@Pi Han Goh Wow, fantastic, good searching. Definitely looks like I took the long way round. It's interesting how powerful half-angle formulas are here (I'm used to using them in calculus problems, but not so much with triangles).

Chris Lewis - 1 month, 4 weeks ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...