Triangle Problem: Max/Min PD+PE+PF

  • Thread starter Thread starter ehrenfest
  • Start date Start date
  • Tags Tags
    Triangle
Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
12 replies · 4K views
ehrenfest
Messages
2,001
Reaction score
1

Homework Statement


A point P is located in the interior of an equilateral triangle ABC. Perpendiculars drawn from P meet each of the sides in points D,E,F, respectively. Where should P be located to make PD + PE + PF a maximum? What about a minimum?


Homework Equations





The Attempt at a Solution



From symmetry considerations. I think that the maximum will be at the incenter. I am not sure about the minimum. Maybe at a vertex, but that is not really in the interior... I tried the standard reflections. I am not sure if I actually need to set up a coordinate system with P at the origin... That looks like it could get messy? I am also not sure how to use the fact that this is an equilateral triangle.
 
Physics news on Phys.org
ehrenfest said:
anyone?

Isn't this one of them Lagrange multiplier things ?

Having said that, i don't know how to set up :

1)the funtion that needs to be minimized/maximized

2) the function that is going to be the constraint.

I would reallly like to see the solution to this problem, though.

Marlon
 
I think it's my english that's bad here, but how can you draw 3 perpendicular lines in a plane?
 
They are not perpendicular to each other.

They are perpendicular to each side of the triangle.
 
How about this approach: (sorry for the double post)

Draw segments from P to each vertex of the triangle. You have split the triangle into three new triangles.

The area of each triangle is given by [tex]\frac{1}{2}sh_i[/tex] where s is the side length of the equilateral triangle and h_i is the height of the perpendicular on that side

Through this, we can set up the equality

[tex]\frac{1}{2}sh_a + \frac{1}{2}sh_b + \frac{1}{2}sh_c = \frac{s^2 \sqrt{3}}{4}[/tex]

[tex]h_a + h_b + h_c = \frac{s \sqrt{3}}{2}[/tex]

meaning that the sum of the three perpendiculars is always constant regardless of the location of P
 
well, take the situation that P is at point a, then if you move D a distance d, for instance, you can still hold E still, but you'll have to move F to keep P one point.

now P moves in the direction of E with a distance: 2d/root[3], this is the distance it shrinks.
the distance it grows is: 2d tan(30) = 2d/root[3], this is equal.

because any motion of P can be described as the sum of the motion in the direction of perpendiculars, the distance is always the same.
 
Last edited:
I must have done something wrong, because chickendude seems right.

EDIT: I've edited my post and saw what i did wrong.
 
Last edited:
ehrenfest;1550359I am not sure if I actually need to set up a coordinate system with P at the origin... That looks like it could get messy? I am also not sure how to use the fact that this is an equilateral triangle.[/QUOTE said:
Thinking about a coordinate system is a good idea, but the choise of te origin is bad :smile:

Let a vertex of the triangle be at the origin, say B and vertex C, lie on [itex]x[/itex] axe. The coordinates of the vertexes are

[tex]A(\frac{\alpha}{2},\frac{\sqrt{3}\,\alpha}{2}, \,B(0,0), \, C(\alpha,0)[/tex]

where [itex]\alpha[/itex] is the side length of the equilateral triangle.
With this arragement the points [tex]E \in AB,\, F \in AC,\, D \in BC[/tex] have coordinates

[tex]E(\frac{\alpha}{4},\frac{\sqrt{3}\,\alpha}{4}),\,F(\frac{3\,\alpha}{4},\frac{\sqrt{3}\,\alpha}{4}),\,D(\frac{\alpha}{2},0)[/tex]

Now let the coordinates of P be [tex](x,y)[/tex]. The function you want to extremize is

[tex]f(x,y)=|PD|+|PE|+|PF|\Rightarrow f(x,y)=\sqrt{(x-\frac{\alpha}{2})^2+y^2}+\sqrt{(x-\frac{\alpha}{4})^2+(y-\frac{\sqrt{3}\,\alpha}{4})^2}+\sqrt{(x-\frac{3\,\alpha}{4})^2+(y-\frac{\sqrt{3}\,\alpha}{4})^2}[/tex]​

By finding the critical points of [tex]f[/tex], i.e. [tex]\vec{\nabla}\,f=0[/tex] you arrive to
  • The minimum value of [tex]f[/tex] correspons to the point [tex]P(\frac{\alpha}{2},\frac{\sqrt{3}\,\alpha}{6})[/tex] which is nothing else but the circumcenter.
  • The maximum value of [tex]f[/tex] corresponds to the vertex A, thus from symmerty every vertex maximizes [tex]f[/tex]. There is no interior point P, which makes [tex]f[/tex] maximum.

Hope I helped! :smile:
 
Rainbow Child said:
By finding the critical points of [tex]f[/tex], i.e. [tex]\vec{\nabla}\,f=0[/tex] you arrive to
  • The minimum value of [tex]f[/tex] correspons to the point [tex]P(\frac{\alpha}{2},\frac{\sqrt{3}\,\alpha}{6})[/tex] which is nothing else but the circumcenter.
  • The maximum value of [tex]f[/tex] corresponds to the vertex A, thus from symmerty every vertex maximizes [tex]f[/tex]. There is no interior point P, which makes [tex]f[/tex] maximum.

Hope I helped! :smile:

Hmmm. I did not see anything wrong with chickendude's post, which means there is probably something wrong here.
 
ehrenfest said:
Hmmm. I did not see anything wrong with chickendude's post, which means there is probably something wrong here.

If you consider the point [itex]p[/itex] to be the tip of three triangles, each of which has one of the sides of the original equilateral triangles as it's base then the total area of those triangles will be:
[tex]\frac{1}{2} l_1 s + \frac{1}{2} l_2 s + \frac{1}{2} l_3 s = \frac{1}{2} (l_1 + l_2 +l_3) s[/tex]
which is also the area of the original triangle, which is clearly constant, so
[tex](l_1+l_2+l_3)[/tex]
must be constant.
 
NateTG said:
If you consider the point [itex]p[/itex] to be the tip of three triangles, each of which has one of the sides of the original equilateral triangles as it's base then the total area of those triangles will be:
[tex]\frac{1}{2} l_1 s + \frac{1}{2} l_2 s + \frac{1}{2} l_3 s = \frac{1}{2} (l_1 + l_2 +l_3) s[/tex]
which is also the area of the original triangle, which is clearly constant, so
[tex](l_1+l_2+l_3)[/tex]
must be constant.

Yes. You just repeated chickendude's proof. I am just saying if that is right, then something is wrong with Rainbow Child's proof.
 
My mistake!

In my proof I took the points

[tex]E(\frac{\alpha}{4},\frac{\sqrt{3}\,\alpha}{4}),\,F (\frac{3\,\alpha}{4},\frac{\sqrt{3}\,\alpha}{4}),\ ,D(\frac{\alpha}{2},0)[/tex]

which are the midpoints of the sides AB, AC and AD! (nothing to do with the original problem :smile:)

So I gave the answer to the question:

"Which point extemizes the sum [tex]|PD|+|PE|+|PF|[/tex] when D, E and F are the midpoints of the sides!"

Sorry!