This solution is from Etienne Chan, age 15, Parramatta Highschool, NSW Australia
I need to prove the area of a triangle A is given by
A = rs
where r is the radius of the incircle (inscribed circle) and
s is the semi-perimeter (half the perimeter (a+b+c)/2).
Let a , b , c be the lengths of the sides of triangle ABC .
Join the incentre I of the triangle to the 3 corners.
From I drop 3 perpendiculars to each of the sides, each of
these has a length of r .
The areas of ABI , BCI , CAI are cr/2 , ar/2 and br/2
respectively. They sum up to give the area of triangle ABC
Area of ABC = cr/2 + ar/2 + br/2 = r(a+b+c)/2 = rs .
Back to the question!
When P is the midpoint of AD, r1 = r2 because they sit on congruent triangles. The area of triangle ABP = AP ×AP/2 = 1/4. PB = √(1+1/4) = √5/2.
The semi perimeter of ABP = (1 + 1/2 + √5/2)/2 = [3+√5]/4.
Using the formula A = rs for the area of the triangle,
the area of ABP = r1[3+√5]/4 = 1/4.
This gives r1 = 1/[3 + √5] = [3 − √5]/4
so r1 = r2 = [3 − √5]/4.
The area of BPC is 1 - 1/4 - 1/4 = 1/2 and
PB = PC = √5/2.
The semi-perimeter of BPC = [1+ √5/2 + √5/2]/2 = [1 + √5]/2.
From the area of triangle BPC we get
r3[1 + √5]/2 = 1/2 so
r3 = 1/[1 + √5] = [√5 − 1]/4.
Now suppose the lengths AP and PD are 1−p and p .
The area of APB = (1−p)/2.
The length PB = √(1 + p2 − 2p + 1) = √(p2 − 2p + 2) and the
semi perimeter of APB = [1 + (1−p) + √(p2 − 2p + 2)]/2.
So, using the area formula again,
|
r2 = |
1−p
2
|
× |
2
[2 − p + √(p2 − 2p + 2)]
|
= |
1
2
|
[2 − p − √(p2 − 2p + 2)]. |
|
Similary with DPC .
The area of DPC = p/2 and the length PC = √(p2 + 1).
The semi perimeter of PC = [1 + p + √(p2 + 1)]/2.
So
|
r1 = |
p
2
|
× |
2
[1 + p + √(p2 + 1)]
|
= |
1
2
|
[1 + p − √(p2 + 1)]. |
|
Similarly
|
r3 = |
1
2
|
× |
2
[1 + √(p2 + 1) + √(p2 −2p + 2)]
|
= |
1
2
|
[1 + √(p2 + 1) − √(p2 −2p +2)][√(p2 +1)− p]. |
|
To finish this question you need to use the formulae found above
which give the radii r1 , r2 and r3 in
terms of p and then consider how these values
change as p varies from 0 to 1. For example r1 varies from 0
to approximately 0.293 as p increases from 0 to 1.
In order to discover whether the ratio of the radii
r1 : r2 : r3 can ever take the value 1 : 2 : 3
you could plot on the same axes the graphs of r1 ,
r2 and r3 as p varies from 0 to 1.