🔢 mathematics

Gradient estimates for $p$ -Laplacian equation with cubic polynomial nonlinearity on Riemannian manifolds

This paper establishes Cheng-Yau type gradient estimates, Liouville theorems, and Harnack inequalities for $p$ -Laplace equations with cubic polynomial nonlinearity on complete Riemannian manifolds with lower Ricci curvature bounds by employing specific transformations and Moser iteration techniques.

Original authors: Zhen Qiu, Youde Wang, Jun Yang

Published 2026-03-03

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Zhen Qiu, Youde Wang, Jun Yang

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are standing on a vast, rolling landscape. This landscape isn't just flat ground; it's a Riemannian manifold, which is a fancy math way of saying "a curved surface that could be shaped like a sphere, a saddle, or something even stranger."

On this landscape, there is a mysterious force field called a solution (let's call it $v$ ). This force field changes as you move across the land. The paper is about understanding how fast this force field can change (its gradient) and what rules govern its behavior.

Here is the breakdown of the paper using simple analogies:

1. The Rules of the Game (The Equation)

The authors are studying a specific rule that the force field must follow. It's a mix of two things:

The "Smoothness" Force ( $p$ -Laplacian): Imagine the force field wants to be as smooth as possible, like water settling in a bowl. The " $p$ " part just changes how "stiff" or "flexible" that smoothing rule is.
The "Push-Pull" Force (Cubic Nonlinearity): This is the unique part of this paper. The force field is being pushed and pulled by three specific "landmarks" on the map, labeled $a_1$ $a_{1}$ , $a_2$ $a_{2}$ , and $a_3$ $a_{3}$ .
- Think of $a_1$ , $a_2$ , and $a_3$ as three distinct valleys or peaks.
- The equation says: "If you are near $a_1$ , you get pushed away. If you are near $a_3$ , you get pulled back. If you are right in the middle at $a_2$ , you are in a delicate balance."

The big question is: If the landscape is curved (like the Earth), how wild can the force field get? Can it change direction instantly? Or does the curvature of the land force it to be gentle?

2. The Problem: The "Middle" Trap

In previous studies, mathematicians had a great tool called a Logarithmic Transformation.

The Analogy: Imagine you are trying to measure the height of a mountain. If the mountain is very tall, you use a logarithmic scale (like the Richter scale for earthquakes) to compress the huge numbers into something manageable.
The Issue: This paper found that this "logarithmic ruler" works perfectly if the force field is stuck between the first two landmarks ( $a_1$ and $a_2$ ) or the last two ( $a_2$ and $a_3$ ).
The Failure: But if the force field wanders all the way from $a_1$ to $a_3$ , passing right through the middle ( $a_2$ ), the logarithmic ruler breaks. It's like trying to measure a mountain that goes both up and down with a ruler that only measures height; the math gets messy and the numbers explode.

3. The Solution: The "Hyperbolic Tangent" Trick

To fix the broken ruler, the authors invented a new tool: the Hyperbolic Tangent Transformation.

The Analogy: Imagine the force field is a rubber band stretched between two cliffs ( $a_1$ and $a_3$ ). The logarithmic ruler was trying to measure the rubber band's tension, but it got confused in the middle.
The new tool is like a magic lens that stretches the rubber band so that the "middle" part ( $a_2$ ) becomes the center of the universe, and the two cliffs ( $a_1$ and $a_3$ ) get pushed infinitely far away.
This transformation turns the messy, wiggly cubic equation into a much cleaner, smoother equation that the mathematicians can actually solve. It's like taking a tangled ball of yarn and using a special comb to straighten it out perfectly.

4. The Main Discovery: The "Speed Limit" (Gradient Estimates)

Once they straightened out the equation, they could prove a Gradient Estimate.

What is a Gradient? It's the "steepness" or "speed" of the change.
The Result: They proved that no matter how curved the landscape is, the force field cannot change too violently. There is a speed limit.
The Formula: The speed limit depends on three things:
1. The dimension of the space (is it a line, a plane, or 3D space?).
2. How curved the landscape is (Ricci curvature).
3. How big the area you are looking at is.
The Takeaway: If the landscape is curved, the force field has to be "gentler." It can't make sharp, jagged turns.

5. The Consequences (Why do we care?)

With this new "speed limit" established, the authors proved two famous things:

The Liouville Theorem (The "No Escape" Rule):
- If the landscape is perfectly flat (no curvature) and the force field is trapped between the landmarks, the authors proved that the force field cannot move at all. It must be a constant, boring, flat line.
- Analogy: If you are in a perfectly flat room with three walls pushing you, and you aren't allowed to speed up, you just have to stand still. You can't wander around.
The Harnack Inequality (The "Neighborhood" Rule):
- This rule says that if you know the value of the force field at one point, you can predict how big or small it can be at a nearby point.
- Analogy: If you know the temperature in your living room is 70°F, and you know the "speed limit" of how fast temperature can change in your house, you know the temperature in the kitchen can't suddenly be 200°F or -50°F. It has to be within a reasonable range.

Summary

This paper is like a group of explorers mapping a strange, curved world. They found that a standard map (logarithmic transformation) failed in the middle of the territory. So, they built a new, better map (hyperbolic tangent transformation). With this new map, they proved that the "weather" (the solution) on this world can't change too violently. They set a speed limit for the weather, which tells us that if the world is flat, the weather must be perfectly calm, and if you know the weather in one spot, you know exactly how wild it can get in the next spot.

1. Problem Statement

The paper investigates the qualitative properties of solutions to a specific class of nonlinear partial differential equations on complete Riemannian manifolds $(M, g)$ with lower Ricci curvature bounds. The primary equation studied is the p-Laplacian equation with cubic polynomial nonlinearity:

$\Delta_p v + (v - a_1)(v - a_2)(v - a_3) = 0$

where:

$\Delta_p v = \text{div}(|\nabla v|^{p-2}\nabla v)$ is the $p$ -Laplacian operator ( $p > 1$ ).
$a_1 < a_2 < a_3$ are real constants.
The manifold has Ricci curvature bounded from below: $\text{Ric}_g \geq -(n-1)\kappa g$ for some $\kappa \geq 0$ .

The authors aim to establish Cheng-Yau type gradient estimates for weak solutions $v$ that lie within specific intervals defined by the constants $a_i$ :

$v \in (a_1, a_2)$
$v \in (a_2, a_3)$
$v \in (a_1, a_3)$ (spanning across the middle root $a_2$ )

Based on these estimates, the paper derives Liouville theorems (conditions under which solutions must be constant) and Harnack inequalities.

2. Methodology

The authors employ a sophisticated combination of geometric analysis techniques, specifically tailored to handle the nonlinearity of the cubic term and the degeneracy of the $p$ -Laplacian.

A. Transformation of Variables

The core strategy involves transforming the original solution $v$ into a new function $w$ to linearize or simplify the nonlinearity. The choice of transformation depends on the interval in which $v$ resides:

Logarithmic Transformation: Used for solutions in $(a_1, a_2)$ $(a_{1}, a_{2})$ and $(a_2, a_3)$ $(a_{2}, a_{3})$ .
- For $a_1 < v < a_2$ : $w = -(p-1)\ln(v - a_1)$ .
- For $a_2 < v < a_3$ : $w = -(p-1)\ln(a_3 - v)$ .
- Rationale: This converts the polynomial nonlinearity into exponential terms, which are convex, facilitating the derivation of lower bounds for the linearized operator.
Hyperbolic Tangent Transformation: Introduced as a novel contribution for solutions in the full interval $(a_1, a_3)$ $(a_{1}, a_{3})$ .
- Transformation: $v = m + L \tanh w$ , where $m = \frac{a_1+a_3}{2}$ and $L = \frac{a_3-a_1}{2}$ .
- Rationale: The standard logarithmic transformation fails when $v$ crosses the middle root $a_2$ because the resulting terms lose convexity. The $\tanh$ transformation maps the infinite domain of $w$ to the finite interval $(a_1, a_3)$ and preserves the necessary structural properties for analysis.

B. Linearized Operator Analysis

The authors analyze the linearization operator $L$ of the $p$ -Laplacian acting on $f = |\nabla w|^2$ . They derive pointwise lower-bound estimates for $L(f^\alpha)$ (where $\alpha$ is a large constant).

The analysis involves completing squares in the nonlinear terms and utilizing the Bochner-type identity for the $p$ -Laplacian.
Crucial to the proof is ensuring that the coefficient of the $f^2$ term remains positive. This imposes specific constraints on the parameter $p$ relative to the dimension $n$ and the constants $a_i$ .

C. Integral Estimates and Moser Iteration

Once the pointwise lower bounds are established, the authors:

Apply Saloff-Coste's Sobolev inequality (adapted for manifolds with lower Ricci curvature bounds).
Use Moser iteration on the integral inequalities derived from the linearized operator.
This iterative process upgrades the local $L^\gamma$ bounds of the gradient to an $L^\infty$ bound, yielding the final gradient estimate.

3. Key Contributions

Novel Transformation for Cubic Nonlinearity: The paper introduces the hyperbolic tangent transformation to handle solutions spanning the entire range $(a_1, a_3)$ . This overcomes the limitation of the standard logarithmic transformation, which is only effective when the solution is confined to one side of the middle root $a_2$ .
Optimal Range of $p$ : The authors derive precise ranges for the exponent $p$ $p$ under which the gradient estimates hold.
- For intervals $(a_1, a_2)$ and $(a_2, a_3)$ : $p \in (\frac{2n+2}{n+3}, 4)$ .
- For the full interval $(a_1, a_3)$ : $p$ must satisfy a more complex condition involving $\delta = \frac{a_1+a_3-2a_2}{a_3-a_1}$ , specifically $p \in [\frac{2n}{n+1}, \frac{2n}{n+1} + \frac{(n-1)(1+\sqrt{1-\delta^2})}{(n+1)\delta^2}]$ .
Unified Framework: The work unifies the treatment of generalized Allen-Cahn and Fisher-KPP type equations (which are special cases of the cubic nonlinearity) under the $p$ -Laplacian framework on Riemannian manifolds.

4. Main Results

A. Cheng-Yau Type Gradient Estimates

For a solution $v$ on a geodesic ball $B_R(o)$ :

Case 1 ( $a_1 < v < a_2$ ):
$\sup_{B_{R/2}(o)} \frac{|\nabla v|^2}{|v - a_1|^2} \leq C(n, p) \left( \frac{1 + \sqrt{\kappa}R}{R} \right)^2$
Case 2 ( $a_2 < v < a_3$ ):
$\sup_{B_{R/2}(o)} \frac{|\nabla v|^2}{|a_3 - v|^2} \leq C(n, p) \left( \frac{1 + \sqrt{\kappa}R}{R} \right)^2$
Case 3 ( $a_1 < v < a_3$ ):
$\sup_{B_{R/2}(o)} \frac{|\nabla v|^2}{4[(v - a_1)(a_3 - v)]^2} \leq \frac{C(n, p)}{(a_3 - a_1)^2} \left( \frac{1 + \sqrt{\kappa}R}{R} \right)^2$

B. Liouville Theorems

On a complete non-compact manifold with non-negative Ricci curvature ( $\kappa = 0$ ):

If $p$ satisfies the conditions for Case 1 or 2, no solution exists in $(a_1, a_2)$ or $(a_2, a_3)$ .
If $p$ satisfies the conditions for Case 3, the only solution in $(a_1, a_3)$ is the constant solution $v \equiv a_2$ .

C. Harnack Inequalities

The gradient estimates imply local Harnack inequalities, controlling the ratio of the solution values at different points within a ball. For example, if $a_1 < v < a_2$ :
$\frac{v(x) - a_1}{v(y) - a_1} \leq e^{C(n, p)(1 + \sqrt{\kappa}R)}$
Similar inequalities are derived for the other intervals, demonstrating the regularity and boundedness of solutions.

5. Significance

Extension of Classical Theory: This work extends the classical Cheng-Yau gradient estimates (originally for harmonic functions, $p=2$ ) and recent results for $p$ -Laplacian equations to the case of cubic polynomial nonlinearities.
Geometric Analysis: It provides a rigorous link between the geometry of the manifold (Ricci curvature) and the analytic properties of nonlinear PDEs, specifically addressing the "gap" in existing literature regarding solutions that cross the unstable equilibrium point ( $a_2$ ) of the cubic potential.
Applications: The results have direct implications for the study of phase transition models (generalized Allen-Cahn equations) and population dynamics (generalized Fisher-KPP equations) on curved spaces, offering new tools to prove non-existence of solutions and regularity properties without assuming positivity of the solution globally.

In summary, the paper successfully generalizes gradient estimate techniques to a broader class of nonlinear $p$ -Laplacian equations by introducing a new transformation method and carefully analyzing the interplay between the nonlinearity, the exponent $p$ , and the manifold's curvature.