I am looking for a reference where the following situation or something similar could have been studied. As a foreword, my question may not be very technical since I am from an engineering background. I have tried to provide an intuitive explanation of the problem. I am also looking for a mildly technical reference or answer:

Heat equation can be used to study diffusion of heat on a surface. On a plane the boundary for a such a heat equation is a circle. I am looking for a system of three heat equation type PDEs (call them $a_{1}$, $a_{2}$ and $a_{3}$) so that some conditions are satisfied.

1) $a_{1}$, $a_{2}$ and $a_{3}$ describe propagation of heat starting at three different points $A_{1}$, $A_{2}$ and $A_{3}$ on the plane.

2) Stopping time of $a_{i}$ is when intersection of boundary of $a_{i}$ and union of boundaries $a_{j}$ and $a_{k}$ is non empty with $i \ne j \ne k \ne i$ $\forall i \in \{1,2,3\}$.

Let the stopping time of $a_{i}$ be $t_{i}$. After giving a simple description of pdes that could satisfy the above conditions, I also need to find expressions for $t_{i}$ which is what I am truly after since it finding $t_{i}$ could provide distance between $A_{i}$ to it closest neighbor without using the euclidean formula. Can one get the PDE to stop diffusing without introducing an artificial stopping time? Can one generalize this to many points in $n$-dimensions (real or complex). Such a system could capture closest neighbors to each given point.

  • $\begingroup$ When you say "the boundary for a such a heat equation is a circle" do you mean the level sets for a fundamental solution are circles? $\endgroup$
    – S. Carnahan
    Apr 25, 2011 at 4:15
  • $\begingroup$ I think that is what I am implying. $\endgroup$
    – Turbo
    Apr 25, 2011 at 5:46

1 Answer 1


This answer makes some assumptions about what the OP is asking. In particular I am using Scott's interpretation that the boundaries of interest correspond to level sets of the fundamental solution.

I suppose you have to specify which level set you are talking about. Since the heat equation has infinite propagation speed, you can make the $t_i$ as close to zero as you like by looking at level sets $a_i = \epsilon$ for $\epsilon$ sufficiently small. If you fix $\epsilon$ (or $\epsilon_i$) then you are looking at a collection of three circles in the plane with radius $r_i(t) = \left(4kt \log(1/\epsilon) + 2kt \log(4 \pi k t)\right)^\frac{1}{2}$ (assuming the standard heat equation with diffusion constant k) and you are asking when these circles intersect. The circle about $A_i$ will intersect the circle about $A_j$ when $r_i + r_j \ge |A_i - A_j|$.

  • $\begingroup$ Hi Aaron: Thank you for the answer. Can you write explicitly the system of PDEs that will start diffusing and stop at the closest neighbor's level set? $\endgroup$
    – Turbo
    Apr 26, 2011 at 1:54
  • $\begingroup$ I am wondering what the system would look like for more than $3$ points in $n \ge 3$ dimensions. $\endgroup$
    – Turbo
    Apr 26, 2011 at 1:55
  • $\begingroup$ Caveat: Maybe I'm not understanding your question. I am imagining $u_t = D\Delta u$ where $u(x,t) \in \R^m$ and $x \in \R^n$ and $D$ is a diagonal matrix with entries $k_i$. This is just a collection of uncoupled heat equations in $\R^n$ whose fundamental solution is just a vector of Gaussians. It sounds like the solution that you are interested in corresponds to $u_i$ a Gaussian centered at $A_i$. The stopping time is introduced artificially by picking levels $\epsilon_i$ and taking logarithms of the Gaussians to solve for the radius of the spherical level set as a function of time ... $\endgroup$ Apr 26, 2011 at 2:24
  • $\begingroup$ (continued) ... and then call a collision any time that $|A_i - A_j| \ge r_i + r_j$. Getting the PDE to stop diffusing without introducing an artificial stopping time would be more subtle. $\endgroup$ Apr 26, 2011 at 2:27
  • $\begingroup$ "Getting the PDE to stop diffusing without introducing an artificial stopping time would be more subtle." Actually I think I am interested in exactly this "automation" of the stopping times so that the description of the PDEs themselves would capture the closest neighbors. Such a description could provide an analytical gadget to some euclidean graph problems $\endgroup$
    – Turbo
    Apr 26, 2011 at 2:46

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.