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

For a nonnegative polynomial $V: \mathbb{R} \to \mathbb{R}$, write $H = -\Delta + V$. I am wondering if there is an elliptic Harnack inequality for H. That is:

There exist $C_{H} > 0$ and $\delta \in (0,1)$ such that for any $B(x,r)$ in $\mathbb{R}$, if $u \geq 0$ satisfies $Hu \equiv 0$ in $B(x,r)$, then

$$\sup_{B(x,\delta r)} u \leq C_{H} \inf_{B(x,\delta r)} u$$

where $C_{H}$ and $\delta$ are indepedent of $x$ and $r$.

I see results such as the "strong Harnack inequality" in Aizenman and Simon's 1982 paper "Brownian Motion and Harnack Inequality for Schrödinger Operators". But their equivalent of the constant $C_{H}$ appears to depend on BOTH $x$ and $r$.

share|cite|improve this question
up vote 3 down vote accepted

The constant $C_H$ must depend on $x$ and $r$. Consider for example $u = e^{\frac{1}{2}|x|^2}$, which satisfies $$\Delta u = (1+|x|^2)u.$$ Then $u$ has a minimum of $1$ at $0$, which clearly doesn't control the maximum of $u$ on $B_{\delta r}$ times any constant independent of $r$.

Furthermore, $C_H$ must depend on $x$ because if we fix $r$ then the ratio $$\frac{u(x+r)}{u(x-r)}$$ blows up as $x \rightarrow \infty$.

share|cite|improve this answer
Thank you, very instructive. – Michael Tinker Jun 1 '13 at 18:27
A Harnack inequality with constants depending on the radius may be found in Aizenman, M.; Simon, B. Brownian motion and Harnack inequality for Schrödinger operators. – Gian Maria Dall'Ara Oct 15 '13 at 7:54

There is a nice presentation of Harnack inequalities for linear elliptic p.d.e in Protter and Weinberger "Maximum Principles in Differential Equations". Moreover, there are additional references to the nonlinear case in the bibliography, which are probably more helpful.

share|cite|improve this answer
I'll check it out! – Michael Tinker Jun 1 '13 at 18:27

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.