Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Consider the p-Laplacian of a suitably nice function $u$:

$\Delta_p u = \nabla \cdot (|\nabla u|^{p-2} \nabla u)$

Are there useful ways of thinking about the p-Laplace operator, or of thinking about p-harmonic functions (the functions $u$ with $\Delta_p u = 0$ in some domain)?

What are some application of the p-Laplacian that could help in building intuition for it?

-

Similar question for the Laplacian: Intuitive interpretation of the Laplacian

share|improve this question
add comment

2 Answers

up vote 6 down vote accepted

One practical application of $\int |\nabla u|^p\to \min$ is in image denoising. If $u$ is the light intensity, then noise contributes to $|\nabla u|$. The easiest thing to do computationally is to decrease $\int |\nabla u|^2$, which more or less amounts to Gaussian smoothing. Unfortunately, $\int |\nabla u|^2$ penalizes not only noise, but also sharp edges, which become blurry in the process.

Taking $p=1$ instead of $p=2$, we arrive at TV denoising, which preserves edges. The drawback here is that the total variation functional is not strictly convex (bad for minimum search) and the associated Euler-Lagrange equation $\nabla \cdot (|\nabla u|^{-1}\nabla u)$ is awfully degenerate. One way around this is to minimize $\int \alpha |\nabla u| + |u-u_0|^2$, with $u_0$ being the noisy image. This addition of the quadratic term makes the optimization process more palatable: see here, for example.

On the opposite extreme, for large $p$, the $p$-Laplacian is used to model growing/collapsing sandpiles.

share|improve this answer
add comment

Well, in a sense that's only useful for other mathematicians, it's about minimizing the $L^p$ norm of the gradient.

In general, you can think of it as another averaging differential operator: it has the form $\nabla \cdot (D \nabla u)$. In heat transport equations, D is the diffusion coefficient, which in this case is proportional to a power of the gradient.

There is also a detailed explanation of some of the physics background of the equation at Math Overflow here

share|improve this answer
add comment

Your Answer

 
discard

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.