MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Some of the classical triangle centers can be expressed as solutions to minimization problems: Given a triangle $A_1, A_2, A_3$ define $d_i, i=1,2,3$ to be the distance of a given point $P$ to $A_i$, and $f_q$ as the sum of the $q$-th power of these distances:$f_q = \sum_{i=1}^3 d_i^q$. I'm looking for the point $P$ which minimizes $f_q$. For $q=1$ this is the Steiner point, for $q=2$ the centroid, for $q \to \infty$ the circumcenter, for $q \to 0$ the point where the product of distances is minimized. An obvious question is to find the curve of all these points for reasonably general $q$ (e.g. $q \in \mathbb{R}_{>0}$). However, in the ressources for triangle centers, as e.g. this problem seems not to be considered.


1) I would like to restrict the problem to $q \ge 1$ since then $f_q$ is convex and a unique minimum is guaranteed.

2) I would like to add a generalization of the question: Consider all continuous functions $f(d_1,d_2,d_3)$ that a) are invariant under permutations of $d_1, d_2, d_3$ and b) have a unique minimum. What can be said about the locus of all these minima?

share|cite|improve this question
Don't you lose convexity when $q$ is small? It seems like you would get local minima near each corner in the zero limit. – S. Carnahan May 30 '10 at 1:41
When $q \rightarrow \infty$, $P$ does not (generally) approach the circumcenter. Consider an obtuse isosceles triangle. If we let the obtuse angle approach $\pi$, the circumcircle radius grows to infinity, so the center grows arbitrarily far from the vertices of the triangle. Such a point clearly cannot minimize $f_q$. In this case, P approaches the midpoint of the side opposite the obtuse angle. – Matthew Conroy Jun 29 '10 at 17:28

I posted an animation showing approximations of this curve for a sequence of triangles, and a range of q values, here.

share|cite|improve this answer
Looks like it's a private video. – j.c. Jun 28 '10 at 13:10
I believe it is now viewable by everyone. – Matthew Conroy Jun 28 '10 at 16:59

Your curve might end up looking quite strange. Consider what happens if $A_1 \rightarrow \infty $ along a ray connecting it to the origin. the Steiner point stays fixed, and the centroid goes to infinity. While I'm not sure, it seems like for any $q > 1$, the point will go to infinity.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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