Parabolic problems whose Fujita critical exponent is not given by scaling

This paper investigates the fractional heat equation with a Riesz potential nonlinearity, establishing a Fujita-type critical exponent that deviates from standard scaling predictions, thereby confirming a hypothesis by Mitidieri and Pohozaev and extending their nonexistence results to general convolution operators.

The Gist

Imagine you are watching a pot of soup on a stove. This soup represents a mathematical "solution" to a complex equation describing how heat or particles spread out over time and space.

In the world of physics and math, there's a famous rule called the Fujita Critical Exponent. Think of this as a "Goldilocks Zone" for the soup:

• If the ingredients (the initial data) are too "hot" or the recipe (the equation) is too aggressive, the soup will explode (blow up) in a finite amount of time.
• If the ingredients are mild enough, the soup will simmer forever without exploding (global existence).

For decades, mathematicians believed they could predict this "explosion point" using a simple rule of thumb called Scaling. Imagine if you doubled the size of your pot and the amount of soup; scaling says the explosion point should shift in a predictable, proportional way.

The Big Twist in This Paper
Authors Ahmad Fino and Berikbol Torebek discovered that for a specific type of "soup" involving Riesz potentials (a fancy way of saying the soup ingredients interact with each other across long distances, not just neighbors touching neighbors), the old rule of thumb fails completely.

Here is a breakdown of their findings using everyday analogies:

1. The "Long-Distance Phone Call" Effect

Usually, in heat equations, a particle only feels the heat of its immediate neighbors. But in this paper, the equation includes a Riesz Potential.

• Analogy: Imagine a party where everyone can hear everyone else, no matter how far apart they are in the room. If one person starts shouting (a high concentration of "u"), everyone in the room feels it instantly. This "long-distance" interaction changes the rules of the game.
• The Result: The point at which the system explodes is not where the old scaling math predicted it would be. It's a new, unexpected "tipping point."

2. The Two Scenarios: Explosion vs. Eternal Simmer

The paper answers three big questions about this "long-distance" soup:

• Question 1: Does the soup explode if the recipe is too spicy?

  • Yes. If the "spiciness" (the power $p$) is below a new critical number ($p_{Fuj}$), the soup will inevitably boil over and explode, no matter how small the initial amount of soup was.
  • The Surprise: This critical number is higher than what the old scaling rules predicted. The system is actually more stable than we thought; it can handle more spice before exploding.

• Question 2: Can we keep the soup simmering forever?

  • Yes, but only if we start small. If the spiciness is above that new critical number, and we start with a very tiny, gentle amount of soup, it will simmer forever without exploding.
  • The Victory: This confirms a guess made by famous mathematicians Mitidieri and Pohozaev years ago. They suspected this was possible, but Fino and Torebek proved it with a rigorous mathematical "recipe."

• Question 3: Does this work for other types of "long-distance" interactions?

  • Yes. They showed that this behavior isn't just about the specific Riesz potential; it applies to a whole family of "convolution" interactions (where ingredients mix in various ways).

3. How Did They Prove It?

To prove the soup would explode, they used a method called the "Test Function" method.

• Analogy: Imagine trying to prove a dam will break. Instead of waiting for the water to rise, you place a series of sensors (test functions) at different spots. You show that if the water level stays below a certain point, the sensors would have to register impossible values (like negative water). Since that's impossible, the dam must break.

To prove the soup could simmer forever, they used a Fixed-Point Argument.

• Analogy: Imagine trying to balance a broom on your hand. If you make tiny, precise adjustments (mathematical iterations), you can keep it upright. They showed that if the initial "wobble" (initial data) is small enough, the mathematical "adjustments" keep the solution stable and bounded forever.

The Takeaway

This paper is a correction to a long-held belief in the physics of heat and diffusion.

• Old Belief: "If you scale the problem up, the explosion point moves predictably."
• New Reality: "When ingredients interact over long distances, the explosion point shifts to a new, higher level. The system is more resilient than we thought, but the math to predict it is more complex."

In short, Fino and Torebek found a new "speed limit" for these equations, showing that the universe of these mathematical models is more nuanced and interesting than the simple scaling rules suggested.

Technical Summary

Here is a detailed technical summary of the paper "Parabolic Problems Whose Fujita Critical Exponent Is Not Given by Scaling" by Ahmad Z. Fino and Berikbol T. Torebek.

1. Problem Statement

The paper investigates the global behavior of solutions to the fractional heat equation with a nonlocal nonlinearity involving a Riesz potential:
$$\begin{cases} u_t + (-\Delta)^{\beta/2} u = I_\alpha(|u|^p), & x \in \mathbb{R}^n, \, t > 0, \\ u(x, 0) = u_0(x), & x \in \mathbb{R}^n, \end{cases}$$
where:

• $\beta \in (0, 2]$ is the order of the fractional Laplacian (with $\beta=2$ corresponding to the classical Laplacian).
• $\alpha \in (0, n)$ is the order of the Riesz potential $I_\alpha$, defined as a convolution with the kernel $|x|^{-(n-\alpha)}$.
• $p > 1$ is the power of the nonlinearity.
• $u_0 \in L^\infty(\mathbb{R}^n)$ is the initial data.

The central objective is to determine the Fujita critical exponent, denoted $p_{Fuj}$, which separates the regime of global existence (for small initial data) from the regime of finite-time blow-up or nonexistence of global solutions.

2. Methodology

The authors employ a combination of analytical techniques tailored to the nonlocal nature of the problem:

• Local Existence and Uniqueness: Established using a Banach fixed-point argument in a suitable function space ($L^\infty((0, T), L^s \cap L^\infty)$). The proof relies heavily on the Hardy–Littlewood–Sobolev (HLS) inequality to handle the Riesz potential term and standard $L^p-L^q$ estimates for the fractional heat semigroup $S_\beta(t)$.
• Blow-up and Nonexistence: Proven using a nonlinear capacity method (also known as the test function method). The authors construct specific test functions $\psi(t, x) = \phi_R^\ell(x)\phi_T^\ell(t)$ adapted to the scaling of the fractional operator and the nonlocal source. By deriving integral inequalities and analyzing the asymptotic behavior as the parameters $R$ and $T$ tend to infinity, they derive contradictions for global solutions when $p$ is below the critical threshold.
• Global Existence: Established via a fixed-point argument in a weighted space involving the decay of the solution. The proof utilizes the HLS inequality and precise decay estimates of the fractional heat kernel to show that for sufficiently small initial data, the solution remains bounded globally in time.
• Generalization: The methodology is extended to a more general convolution operator $(K * |u|^p)$ where $K$ is a kernel satisfying specific growth/decay conditions, not limited to the Riesz potential.

3. Key Contributions and Results

A. The Critical Exponent

The most significant finding is the identification of the critical Fujita exponent:
$$p_{Fuj}(n, \beta, \alpha) = 1 + \frac{\beta + \alpha}{n - \alpha}$$
Crucial Distinction: The authors demonstrate that this exponent is not determined by the standard scaling argument.

• The scaling-invariant exponent (derived from the invariance of the equation under $u_\lambda(t, x) = \lambda^{\frac{\beta+\alpha}{p-1}}u(\lambda^\beta t, \lambda x)$) is:
  $$p_{sc} = 1 + \frac{\beta + \alpha}{n}$$
• Since $n - \alpha < n$, it follows that $p_{Fuj} > p_{sc}$.
• This implies that the "usual" scaling argument fails to predict the blow-up threshold for this class of nonlocal problems. The critical exponent is driven by the interplay between the fractional diffusion and the long-range interaction of the Riesz potential, rather than just dimensional scaling.

B. Main Theorems

1. Local Existence (Theorem 1.2): A unique mild solution exists locally in time for $p > n/(n-\alpha)$.
2. Blow-up (Theorem 1.4(i)): If $u_0 \geq 0$, $u_0 \in L^1 \cap L^\infty$, $\int u_0 > 0$, and $1 < p \leq p_{Fuj}(n, \beta, \alpha)$, the solution blows up in finite time.
3. Global Existence (Theorem 1.4(ii)): If $p > p_{Fuj}(n, \beta, \alpha)$ and $\|u_0\|_{L^{q_{sc}}}$ is sufficiently small (where $q_{sc} = n(p-1)/(\beta+\alpha)$), a global solution exists.
4. General Kernel Nonexistence (Theorem 1.7 & 1.9): The blow-up/nonexistence results are extended to equations with a general convolution kernel $K$. Specifically, if the kernel decays slowly enough (related to the Riesz potential case), nonexistence of global weak solutions holds under similar conditions.

C. Resolution of Open Questions

The paper answers three specific questions posed in the introduction:

1. Mitidieri–Pohozaev Conjecture: The authors confirm the conjecture that for $\beta=2$, if $p > 1 + (2+\alpha)/(n-\alpha)$, global solutions exist for small initial data. This provides a positive answer to the hypothesis proposed in [24].
2. Blow-up vs. Nonexistence: They rigorously establish that the nonexistence of global solutions for $p \leq p_{Fuj}$ corresponds to a finite-time blow-up phenomenon for mild solutions, clarifying the distinction between weak nonexistence and strong blow-up.
3. Generalization: They successfully extend the results from the specific Riesz potential to a broader class of convolution operators.

4. Significance

• Theoretical Breakthrough: The paper challenges the conventional wisdom that scaling arguments always dictate the Fujita exponent. It highlights a unique class of parabolic problems where the nonlocal source term fundamentally alters the critical threshold, making it strictly larger than the scaling prediction.
• Unification of Results: It unifies and extends previous work on classical heat equations ($\beta=2, \alpha=0$), fractional heat equations ($\alpha=0$), and heat equations with nonlocal nonlinearities (Cazenave et al., Mitidieri & Pohozaev).
• Methodological Rigor: The adaptation of the test function method to handle the specific structure of the Riesz potential and fractional Laplacian provides a robust framework for analyzing similar nonlocal evolution equations.
• Physical Relevance: The results are relevant for modeling physical phenomena involving long-range interactions (modeled by Riesz potentials) and anomalous diffusion (modeled by fractional Laplacians), such as in plasma physics, fluid dynamics, and population dynamics.

In summary, Fino and Torebek provide a comprehensive analysis of a fractional parabolic equation with nonlocal nonlinearity, proving that the critical exponent is governed by a complex balance of diffusion and nonlocal interaction rather than simple scaling, and rigorously establishing the conditions for global existence and blow-up.