The heat equation and independence of the spectrum of the Hodge Laplacian on $\ell^p$

This paper establishes Davies-Gaffney-Grigoryan estimates for the heat kernel of the Hodge Laplacian on simplicial complexes, extends the associated semigroup to $\ell^p$ spaces under specific curvature and volume growth conditions, and proves the independence of the spectrum from $p$ for $p \in [1, \infty]$.

The Gist

Imagine you have a giant, complex Lego structure. It's not just a flat picture; it's a 3D web of shapes, connections, and layers. In mathematics, this is called a simplicial complex. Now, imagine you drop a single drop of hot dye onto one specific Lego brick. Over time, that dye spreads out, mixing with its neighbors, flowing through the connections, and eventually coloring the whole structure.

This spreading process is the Heat Equation.

The paper you're asking about is a sophisticated study of how this "heat" (or information, or probability) moves across these complex Lego structures. But instead of just watching it happen, the authors are asking some very deep questions:

1. How fast does it spread?
2. Does the way we measure the "amount" of dye change how the heat behaves?
3. Does the shape of the Lego structure (its geometry) dictate the rules of the spread?

Here is a breakdown of their findings using everyday analogies.

1. The Main Characters: The Hodge Laplacian and the Heat

In the world of smooth surfaces (like a sphere or a donut), mathematicians use a tool called the Hodge Laplacian to describe how things vibrate or diffuse. Think of it as the "rulebook" for how heat moves.

In this paper, the authors look at discrete structures (like our Lego set) instead of smooth surfaces. They want to know: If we have a rulebook for heat on a Lego set, does it behave the same way whether we count the dye in "buckets" (standard math) or "squares" (different math)?

2. The Big Problem: The "Perspective" Issue ($\ell^p$)

Mathematicians measure things in different ways, called $\ell^p$ spaces.

• $\ell^2$ (The Standard View): This is like measuring the total energy of the heat. It's the most common way to look at things, like taking a photo of the whole Lego set.
• $\ell^1$ and $\ell^\infty$ (The Extreme Views): These are like looking at the total amount of dye (summing everything up) or looking at the single hottest spot (the maximum).

The Question: If the heat spreads nicely in the "Standard View" ($\ell^2$), does it also spread nicely in the "Extreme Views" ($\ell^1$ or $\ell^\infty$)? Or does the math break down if we change our perspective?

For a long time, mathematicians knew the answer for simple, flat grids (graphs). But for complex, multi-layered Lego structures (simplicial complexes), it was a mystery.

3. The Solution: The "Heat Map" Estimate

The authors developed a new way to estimate how fast the heat kernel (the "dye drop") spreads. They call this a Davies-Gaffney-Grigoryan estimate.

The Analogy: Imagine you are in a dark forest (the Lego structure). You drop a flare. How far can the light reach after 1 second?

• If the forest is dense and full of obstacles, the light doesn't go far.
• If the forest is open, it goes far.

The authors proved a formula that says: "No matter how weird the Lego structure is, the heat cannot travel faster than a certain speed, and it fades away exponentially based on the distance."

This formula is the key. It acts like a safety net. Once they proved this safety net exists, they could show that the heat equation works perfectly well, even when we switch from the "Standard View" ($\ell^2$) to the "Extreme Views" ($\ell^1$ or $\ell^\infty$).

4. The Two Rules of the Game

To make this work, the Lego structure has to follow two specific rules:

Rule A: The "Curvature" Rule (Form Bounded Curvature)
Imagine the Lego structure is a bumpy landscape. Some parts are steep hills (high curvature), some are deep valleys (negative curvature).

• The authors found that as long as the "valleys" aren't too deep relative to the "hills," the heat behaves well.
• They call this Form Bounded Curvature. It's like saying, "The ground can be bumpy, but it can't have a bottomless pit that swallows the heat."

Rule B: The "Growth" Rule (Subexponential Volume Growth)
Imagine you start at one Lego brick and count how many bricks are within 1 step, 2 steps, 3 steps, etc.

• If the number of bricks explodes exponentially (1, 10, 100, 1000...), the structure is too "spacious" and the heat might get lost.
• The authors require Subexponential Growth. This means the structure can grow, but not too fast. It's like a city that expands, but not so fast that you can't find your way around.

5. The Big Discovery: The Spectrum is Independent

The "Spectrum" is like the fingerprint of the Lego structure. It tells you the natural frequencies at which the structure vibrates. If you pluck a guitar string, the notes it makes are its spectrum.

The Mind-Blowing Result:
The authors proved that the fingerprint is the same, no matter how you measure it.

• Whether you measure the heat using the "Standard View" ($\ell^2$) or the "Extreme Views" ($\ell^1$ or $\ell^\infty$), the set of natural frequencies (the spectrum) remains identical.

Why is this cool?
Usually, changing the way you measure a system changes the results. Imagine measuring a room's temperature with a thermometer vs. a thermal camera; sometimes they give different data. But here, the authors showed that for these complex structures, the "truth" of the system is robust. It doesn't matter if you look at the big picture or the tiny details; the fundamental nature of the structure stays the same.

6. Special Case: The "Combinatorial" Lego Set

The paper also looks at a specific type of Lego set where every piece is identical (standard weights).

• They found that if the structure doesn't grow too fast (subexponential), you don't even need to worry about the "bumpy hills" (curvature) at all!
• In this simple case, the heat equation works perfectly, and the spectrum is always the same, regardless of how you measure it.

Summary

Think of this paper as a master plumber fixing the pipes in a massive, multi-story building (the simplicial complex).

1. They figured out exactly how water (heat) flows through the pipes, even if the building is weirdly shaped.
2. They proved that as long as the building isn't growing infinitely fast and the pipes aren't clogged with deep holes, the water flows smoothly.
3. Most importantly, they proved that the water pressure (the spectrum) is the same whether you measure it on the ground floor or the top floor.

This is a huge step forward because it connects the messy, complex world of discrete shapes (like computer networks or social graphs) with the elegant, smooth world of classical geometry, showing that the fundamental laws of physics (heat and vibration) hold true in both.

Technical Summary

1. Problem Statement and Motivation

The paper addresses the extension of the theory of the Hodge Laplacian from the classical $L^2$ setting (Hilbert spaces) to general $L^p$ spaces ($1 \le p \le \infty$) on discrete structures, specifically simplicial complexes.

• Context: In Riemannian geometry, the behavior of the heat semigroup and the spectrum of the Laplacian on $L^p$ spaces is well-studied, revealing phenomena like $L^p$-cohomology and spectral independence. In the discrete setting, while the $L^2$ theory for graphs and simplicial complexes is established (notably by Dodziuk), the $L^p$ theory remains largely unexplored, particularly for infinite complexes and higher-dimensional forms.
• Core Questions:
  1. Under what geometric conditions (curvature, volume growth) does the heat semigroup generated by the Hodge Laplacian on $\ell^2$ extend to a strongly continuous semigroup on $\ell^p$?
  2. Does the spectrum of the Hodge Laplacian depend on $p$? (i.e., Is $\sigma(\Delta_p) = \sigma(\Delta_2)$?)
  3. Can these results be established under weak assumptions (e.g., form-bounded curvature rather than uniform lower bounds) for general weighted simplicial complexes?

2. Methodology and Framework

The authors employ a unified approach by representing the Hodge Laplacian on simplicial complexes as a magnetic Schrödinger operator on a graph. This allows them to leverage techniques from the theory of Schrödinger operators.

• Representation: They utilize the framework from their previous work [BK25] to show that the Hodge Laplacian $\Delta_H = \delta\partial + \partial\delta$ can be written as a signed magnetic Schrödinger operator $H = \Delta$.
  • The operator acts on functions (representing forms) on the complex.
  • The "potential" term in the Schrödinger operator corresponds to the Forman curvature ($c_H$).
• Key Tools:
  • Davies-Gaffney-Grigoryan Estimates: They derive off-diagonal decay estimates for the heat kernel using intrinsic metrics.
  • Form Boundedness: Instead of requiring uniform lower bounds on curvature, they assume the negative part of the potential (curvature) is form-bounded with respect to the quadratic form of the Laplacian.
  • Semigroup Extension: They use interpolation, duality, and specific decay estimates to extend the semigroup from $\ell^2$ to $\ell^p$.
  • Analyticity and Resolvents: They analyze the analyticity of the semigroups and the properties of the resolvent operators to prove spectral independence.

3. Key Contributions and Results

A. Davies-Gaffney-Grigoryan Estimates (Section 2)

The authors prove a generalized Davies-Gaffney-Grigoryan lemma for magnetic Schrödinger operators on graphs with an intrinsic metric $d$ and finite jump size $s$.

• Result: They establish an off-diagonal decay estimate for the heat kernel $p_t(x,y)$:
  $$|p_t(x,y)| \le \frac{1}{\sqrt{m(x)m(y)}} \exp\left(-\lambda_2 t - \frac{t}{s^2}\zeta\left(\frac{s d(x,y)}{t}\right)\right)$$
  where $\lambda_2$ is the bottom of the spectrum and $\zeta$ is a specific function involving $\text{arcsinh}$.
• Significance: This generalizes previous results for bounded Laplacians to unbounded operators and magnetic potentials, providing the necessary kernel bounds to control the semigroup on $\ell^p$.

B. Extension to $\ell^p$ Spaces (Section 3)

The paper establishes two distinct criteria under which the heat semigroup extends from $\ell^2$ to $\ell^p$:

1. Form-Bounded Curvature (Theorem 3.1):

   • If the negative part of the potential (curvature) is form-bounded with bound $M$, the semigroup extends to $\ell^p$ for $p$ in a specific interval $I_M$ centered at 2.
   • If the curvature is bounded from below ($M=0$), the extension holds for all $p \in [1, \infty]$.
   • This allows solving the heat equation $(HE)$ for initial data in these $\ell^p$ spaces.

2. Volume Growth Conditions (Theorem 3.6):

   • If the graph has exponential volume growth, the semigroup extends to $\ell^p$ for all $p \in [1, \infty)$.
   • If the graph has subexponential volume growth, the extension holds for all $p \in [1, \infty]$.
   • This result relies on the interplay between the intrinsic metric, the jump size, and the volume of metric balls.

C. Independence of the Spectrum (Section 4)

The central spectral result is the $p$-independence of the spectrum for the Hodge Laplacian.

• Theorem 4.4: If the potential is form-bounded and the graph has uniform subexponential volume growth, then:
  $$\sigma(\Delta_p) = \sigma(\Delta_2) \quad \text{for all } p \in [1, \infty].$$
• Methodology: The proof involves showing that the resolvent $(\Delta_2 - z)^{-1}$ extends to a bounded operator on $\ell^p$ for $z$ in the resolvent set of $\Delta_2$. This is achieved by:
  1. Proving the resolvent kernel decays exponentially (Lemma 4.7).
  2. Using a conjugation technique with Lipschitz functions (Lemma 4.6) to bound the resolvent on weighted spaces.
  3. Applying analytic continuation arguments to show the resolvent sets coincide.

D. Application to Simplicial Complexes (Section 5)

The authors apply the general Schrödinger operator results to weighted simplicial complexes.

• Form Bounded Curvature: They define the Forman curvature $c_H$ and show that the condition of form-bounded curvature is natural and sufficient for the $L^p$ extension.
• Combinatorial Weights ($m \equiv 1$):
  • They prove that for combinatorial complexes, subexponential volume growth with respect to the standard graph metric implies the existence of an intrinsic metric with finite jump size.
  • Theorem 1.1 & 1.2: They confirm that for combinatorial simplicial complexes:
    • The heat equation is solvable on $\ell^p$ under form-bounded curvature or exponential growth.
    • The spectrum is independent of $p$ under subexponential volume growth (and no curvature assumption is needed if the Laplacian is bounded, which occurs in finite-dimensional combinatorial settings with bounded degree).

4. Significance and Impact

• Bridging Continuous and Discrete: The paper successfully translates deep results from Riemannian geometry (spectral independence, $L^p$-theory) to the discrete setting of simplicial complexes, filling a significant gap in the literature.
• Weakening Assumptions: Unlike previous works that required uniform lower bounds on curvature or bounded Laplacians, this paper works under form-bounded curvature and subexponential volume growth, which are much weaker and more natural geometric conditions.
• Unified Framework: By treating the Hodge Laplacian as a magnetic Schrödinger operator, the authors provide a powerful, unified framework that applies to various discrete Laplacians (combinatorial, magnetic, etc.).
• Spectral Stability: The proof of $p$-independence of the spectrum is a major theoretical advance, ensuring that spectral invariants derived from the Hodge Laplacian are robust regardless of the function space ($\ell^p$) used, provided the geometric growth conditions are met.

In summary, this paper establishes a rigorous foundation for the $L^p$-theory of the Hodge Laplacian on infinite simplicial complexes, proving that the heat equation is well-posed and the spectrum is stable under broad geometric conditions, thereby generalizing classical results from manifolds to discrete geometry.