Why are the Partial Differential Equations so named? i.e, elliptical, hyperbolic, and parabolic. I do know the condition at which a general second order partial differential equation becomes these, but I don't understand why they are so named?

Does it has anything to do with the ellipse, hyperbolas and parabolas?


A general 2nd order linear PDE in two variables is written

$$Au_{xx} + 2Bu_{xy} + Cu_{yy} + Du_x + Eu_y + F = 0$$

and $A,B,C,D,E,F$ can be functions depending on $x$ and $y$. We say a PDE is elliptic, hyperbolic or parabolic if \begin{align} B^2 - AC &= 0, &\text{parabolic} \\ B^2 - AC &>0, &\text{hyperbolic} \\ B^2 - AC &<0, &\text{elliptic} \end{align} Note that if $A,B,C,D,E,F$ depend on $x$ or $y$, there can be regions where the PDE is elliptic, hyperbolic or parabolic and different techniques are used to solve each type. If the coefficients are constant the naming comes form considering the polynomial equation $$Ax^2 + 2Bxy + Cy^2 + Dx + Ey + F = 0$$ depending on the sign of $B^2 - AC$, this forms an ellipse, hyperbola or parabola in $\mathbb{R}^2$. This can be extended to higher dimensions as well with hyperboloids, paraboloids, or ellipsoids.

  • $\begingroup$ Thank You :) , But I think the solutions of the corresponding differential equations have nothing to do with their name? Or do they have?. $\endgroup$ – The_Lazy_Panda Apr 30 '20 at 8:50
  • 3
    $\begingroup$ @VivekanandMohapatra actually, the solutions to simple elliptical PDEs around a small pertubation tend to come out as “blobs”, ellipse-ish, to parabolic PDEs they disperse ever slower like the arms of a parabola, and for hyperbolic they wander off asymptotically straight towards infinity like a hyperbola. But I'm not sure if there's really a rigorous mathematical aspect to this. $\endgroup$ – leftaroundabout Apr 30 '20 at 15:32
  • 1
    $\begingroup$ @leftaroundabout could you give an explicit example of some of these? I don't quite understand what you could consider a hyperbola-like blob in a solution of the hyperbolic wave equation, for example. $\endgroup$ – Ruslan May 1 '20 at 5:06
  • 1
    $\begingroup$ @Ruslan if you start with $u_0(x) = 2\cdot e^{-x^2}$, then $u(x,t) = e^{-(x-t)^2} + e^{-(x+t)^2}$. On a colour plot of this in the $t\geq0$ half-plane, you see something V-shaped with straight arms and rounded near 0, which is also what a $t^2 - x^2 = c$ hyperbola looks like. $\endgroup$ – leftaroundabout May 1 '20 at 9:11
  • 1
    $\begingroup$ @leftaroundabout There is a large rigorous mathematical theory of that kind of things: microlocal analysis. It allows to generalize and understand the concept of ellipticity (or lack thereof) much further. $\endgroup$ – Joonas Ilmavirta May 1 '20 at 11:54

All quadratic curves can be studied using the equation $Ax^2+2Bxy+Cy^2 + Dx + Ey + F=0$ the discriminant of which is $B^2-AC$ and the solution curve will be a ellipse, hyperbola, or parabola depending on whether the discriminant is positive, negative, or zero.

Partial second-order differential equations take a much similar form with $Au_{xx}+2Bu_{xy}+Cu_{yy} + Du_x + Eu_y + F = 0$ and the discriminant plays a similar role in classifying the solutions so they are named after the algebraic curves that resemble this expression.

  • 5
    $\begingroup$ I thought using 2Bxy rather than Bxy made the discriminant B^2-AC, not B^2-4AC? $\endgroup$ – No Name Apr 30 '20 at 2:23

I'd like to address this comment of the OP to one of the answers:

But I think the solutions of the corresponding differential equations have nothing to do with their name? Or do they have?.

Actually, we can relate the names to the corresponding geometrical curves. To do this, consider how the following homogeneous PDE will look in the Fourier-transformed form.

Original PDE (with $u^{(n,m)}(x,y)$ denoting $n$th partial derivative of $u$ in $x$ and $m$th in $y$):

$$Au^{(2,0)}(x,y) + 2Bu^{(1,1)}(x,y) + Cu^{(0,2)}(x,y) + Du^{(1,0)}(x,y) + Eu^{(0,1)}(x,y) = 0.\tag1$$

Fourier-transformed one (with $\hat u(k_x,k_y)$ denoting the Fourier transform of $u(x,y)$):

$$\mathcal L\hat u(k_x,k_y) = 0,\tag2$$


$$\mathcal L=Ak_x^2 + 2Bk_xk_y + Ck_y^2 + Dk_x + Ek_y.\tag3$$

This multiplication by $\mathcal L$ is the Fourier-space version of the differential operator from $(1)$. Notice that $\mathcal L$ is just a second-degree polynomial in $k_x$, $k_y$ — a quadratic form. Now consider the nodes of $\mathcal L$ in the $(k_x,k_y)$ plane: they will be defined by the equation

$\mathcal L=0.\tag4$

Geometrically, these nodes will be exactly the kinds of curves with the names corresponding to the name of the kind of PDE, i.e. ellipses, hyperbolas or parabolas.

  • $\begingroup$ This is a very nice answer, and I think is aimed more towards what the OP was asking. Further, @leftaroundabout's response to that comment in that small perturbations also seem to geometrically represent such equations, is in line with describing the perturbations as proportional to the fourier transform of the equation. I'm curious how to say this in a rigorous way, let me know if this makes some sense. $\endgroup$ – Jon Snow Dec 2 '20 at 4:16

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.