I am asked to find, using Lagrange multipliers, the position of the vertices of a right triangle inscribed on $x^2+4y^2=1$ that has the maximum area. The two legs of the triangle (which are not the hypothenuse) are parallel to the elypsis' axis.

What I have so far is:

$$ 2y = 2 \lambda x\\ 2x = 8 \lambda y\\ x^2+4y^2=1 $$

By solving these equations it is known that both $\lambda$ and $x$ cannot be zero. I got minimum points to be:

$$ f \left( \mp \frac{1}{\sqrt{2}} , \pm \frac{1}{2\sqrt{2}} \right) = \frac{1}{2} $$

Since the textbook has no answer, I want to make sure if that is correct.


Let's say that we need to find a rectangle given the same conditions. Are the critical points the same (which would lead to an area of $1$)?


Area - Geogebra


  • $\begingroup$ Is one vertex the origin and the two (short) legs parallel to the $x$ and $y$ axes? $\endgroup$ – David G. Stork May 26 '16 at 23:58
  • $\begingroup$ @DavidG.Stork I just edited the question David: the two catheti are parallel to the axis of the elypsis. Thank you. $\endgroup$ – bru1987 May 27 '16 at 0:00
  • $\begingroup$ Whew... that makes the problem much simpler. $\endgroup$ – David G. Stork May 27 '16 at 0:03
  • $\begingroup$ Hahahahh yes, sorry about that! $\endgroup$ – bru1987 May 27 '16 at 0:04

You wish to maximize $A = f(x,y) = {1 \over 2} (2 x) (2 y) = 2 x y$ subject to the constraint $g(x,y) = x^2 + 4 y^2 - 1 = 0$.

Form the Lagrangian:

${\cal L}(x,y,\lambda) = f(x, y) - \lambda g(x,y) = 2 x y - \lambda (x^2 + 4 y^2 - 1)$.

Take the three derivatives and set them to zero:

  • ${\partial {\cal L}(x,y,\lambda) \over \partial x} = 2 y - 2 \lambda x = 0$
  • ${\partial {\cal L}(x, y, \lambda) \over \partial y} = 2 x - \lambda 8 y = 0$
  • ${\partial {\cal L}(x, y, \lambda) \over \partial \lambda} = -(x^2 + 4 y^2 - 1) = 0$

Then solve these three simultaneous equations for $x,y, \lambda$, then plug in $x$ and $y$ to the formula for area.

I get $x = -{1 \over \sqrt{2}}, y = -{1 \over 2 \sqrt{2}}$ and thus $A = {1 \over 2}$.

enter image description here

If the problem is instead to maximize the area of a rectangle, then the area function is doubled: $f(x,y) = (2 x)(2 y)$ and the optimum turns out to be the same geometry (and of course twice the area), even though the solution $\lambda$ differs from the triangle case.

  • $\begingroup$ would you mind double checking your partial derivatives? I seem to have found $2 \times$ what you've got for all of them... $\endgroup$ – bru1987 May 27 '16 at 0:22
  • $\begingroup$ Not really David. Let me check everything again... $\endgroup$ – bru1987 May 27 '16 at 0:24
  • $\begingroup$ David would you mind checking the image I uploaded? It seems that it is actually $1/2$, agreed? $\endgroup$ – bru1987 May 27 '16 at 0:35
  • $\begingroup$ I didn't answer "yes" to that as you may check on the discussion above. But nevermind, it was just a misunderstanding. Thank you anyways. $\endgroup$ – bru1987 May 27 '16 at 0:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.