Inner Lipschitz approximation in o-minimal structures

This paper demonstrates that in o-minimal structures, every definable mapping that is Lipschitz with respect to the inner metric can be approximated by $\mathscr{C}^1$ (or $\mathscr{C}^\infty$ if the structure admits smooth cell decomposition) mappings with arbitrarily close derivative bounds, a result achieved through the construction of partitions of unity with sharp derivative estimates and extended to outer Lipschitz mappings.

The Gist

Imagine you are an architect trying to renovate a building that has been damaged by an earthquake. The building is made of strange, jagged materials (this represents a "singular space" in mathematics). Your goal is to smooth out the rough edges and make the walls perfectly straight and smooth (this represents making a function "smooth" or $C^\infty$), but you have a strict rule: You cannot stretch or compress the building too much during the renovation. If you stretch a wall too far, the whole structure might collapse.

This paper is about a team of mathematicians (Nguyen, A. Valette, and G. Valette) who figured out how to perform this renovation on "weird" mathematical shapes while keeping the stretching under control.

Here is a breakdown of their work using everyday analogies:

1. The Problem: The "Inner" vs. "Outer" Distance

Usually, when we measure distance between two points, we draw a straight line between them (like a crow flying). This is the Euclidean distance.

However, imagine the building is a maze with walls. If you want to walk from point A to point B, you can't fly; you have to walk along the corridors. The shortest path inside the maze is the Inner Distance.

• The Challenge: Some mathematical shapes are so jagged that the "straight line" distance is very different from the "walking path" distance.
• The Goal: The authors want to smooth out functions (mathematical rules that tell you how to move from one point to another) that are defined on these jagged shapes. They want to make these rules smooth (like a gentle curve) without making the "walking path" distance change too much. This is called Inner Lipschitz Approximation.

2. The Obstacle: The "Rigid" Rules

In the world of these mathematical structures (called o-minimal structures), there is a catch. These structures are very "rigid."

• The Analogy: Imagine you are trying to smooth a piece of clay, but the clay is made of a material that, if you try to make it perfectly smooth in one spot, it forces the entire connected piece to become smooth or zero out completely. It's like a magical clay that refuses to have a "bump" unless the whole thing is a bump.
• The Result: This makes it very hard to create "smooth partitions of unity." In math, a "partition of unity" is like a set of overlapping spotlights that cover a whole stage. You want to turn the lights on and off smoothly to blend different parts of the stage together. In these rigid structures, turning a light on smoothly often breaks the rules of the structure.

3. The Solution: The "Smart Flashlight" (Partitions of Unity)

To fix this, the authors had to invent a new kind of "spotlight" (a mathematical tool called a partition of unity).

• The Innovation: They built a spotlight that can turn on and off very smoothly, but with a special trick: The brightness changes very slowly.
• Why it matters: If you turn a light on too quickly, it creates a "sharp edge" (a high derivative), which violates the "no stretching" rule. By making the light fade in and out very gradually (with "sharp bounds for the derivative"), they can blend the jagged parts of the shape into smooth parts without stretching the fabric of the shape too much.

4. The Main Achievement: Smoothing the Rough Edges

The paper proves two main things:

1. The Basic Fix: If you have a jagged, "inner Lipschitz" map (a rule that doesn't stretch the walking path too much), you can replace it with a smooth ($C^1$) rule that is almost identical, and the "stretching" (the derivative) will be almost exactly the same as the original.
2. The Super Fix: If the mathematical structure allows for it (has "C-infinity cell decomposition"), you can make the rule infinitely smooth ($C^\infty$). This is the mathematical equivalent of polishing the building until it shines like glass, while still obeying the rule that you didn't stretch the walls.

5. Why This Matters

Why do we care about smoothing jagged shapes without stretching them?

• Real-world Application: This is crucial for solving Partial Differential Equations (PDEs). Think of PDEs as the laws of physics (like how heat flows or how a drum vibrates).
• The Scenario: Imagine trying to calculate how heat flows through a broken, jagged rock. The rock has cracks and sharp corners. Standard math tools often fail on these jagged edges.
• The Benefit: This paper provides a way to "smooth out" the rock just enough to use standard physics tools, while guaranteeing that the "heat flow" (the solution) isn't distorted by the smoothing process. It allows mathematicians to study complex, broken shapes using the powerful tools of smooth calculus.

Summary in a Nutshell

The authors found a way to smooth out jagged mathematical shapes without stretching the fabric of the shape too much. They did this by inventing a special kind of mathematical "fading light" that blends rough edges into smooth curves. This allows scientists to apply smooth, easy-to-use math tools to complex, broken, or irregular shapes, which is essential for solving difficult problems in physics and engineering.

They also dedicated this work to David Trotman, a mathematician who was a giant in the field of understanding how shapes behave near their rough edges.

Technical Summary

Here is a detailed technical summary of the paper "Inner Lipschitz Approximation in O-Minimal Structures" by Nhan Nguyen, Anna Valette, and Guillaume Valette.

1. Problem Statement

The paper addresses the problem of approximating definable Lipschitz mappings within the framework of o-minimal structures. Specifically, it focuses on two distinct types of Lipschitz conditions:

• Outer Lipschitz: Mappings bounded by the standard Euclidean distance ($|f(x) - f(y)| \leq L|x-y|$).
• Inner Lipschitz: Mappings bounded by the inner metric ($\mathring{d}_A$), which is the infimum of the lengths of definable arcs connecting two points within the set $A$.

The Core Challenge:
While it is known that definable Lipschitz functions can be approximated by $C^1$ (or $C^\infty$) definable functions, preserving the Lipschitz constant (or the bound of the derivative) is difficult, especially on singular spaces (sets that are not manifolds).

• In o-minimal structures, particularly those that are polynomially bounded (e.g., semi-algebraic sets), constructing $C^\infty$ definable partitions of unity is generally impossible due to rigidity (a $C^\infty$ definable function with a zero Taylor series at a point must be zero on the connected component).
• The authors aim to show that Inner Lipschitz mappings (which naturally arise in analysis on singular varieties) can be approximated by $C^1$ (and $C^\infty$ where possible) definable mappings with arbitrarily close bounds on their derivatives.

2. Methodology

The authors employ a combination of o-minimal geometry, stratification theory, and constructive analysis.

A. Framework and Definitions

• O-minimal Structures: The work is set in a fixed polynomially bounded o-minimal structure expanding $(\mathbb{R}, +, \cdot)$.
• Inner Metric ($\mathring{d}_A$): Defined as the infimum of lengths of definable $C^0$ arcs. The paper proves (Lemma 3.3) that this metric is equivalent to the infimum over all continuous arcs, ensuring regularity.
• Stratifications: The authors utilize $(w)$-regular stratifications (Kuo-Verdier condition), which imply Whitney's $(b)$-regularity. This allows for the control of angles between tangent spaces of different strata.
• Rugose Vector Fields: They utilize the extension property of vector fields tangent to strata, which are "rugose" (satisfy a Lipschitz-like condition across strata).

B. Key Technical Tools

1. Sharp Partitions of Unity (Theorem 4.2):

   • The authors construct definable $C^1$ partitions of unity subordinate to a stratification.
   • Crucial Innovation: Unlike standard partitions of unity, these are constructed with sharp derivative bounds. Specifically, for a stratum $S$, the gradient of the partition function $\phi_S$ satisfies $|\nabla \phi_S(x)| \leq \frac{\mu}{d(x, S)}$. This bound is critical for controlling the derivative of the approximated function near singularities.
   • The construction relies on a specific "bump function" $\psi_\mu$ (Lemma 4.1) with controlled partial derivatives.

2. Inner Lipschitz Characterization (Lemma 3.2):

   • They establish that the local inner Lipschitz constant $\mathring{L}f(x_0)$ is equal to the infimum of the supremum of the derivative norm $|df|$ in a neighborhood. This bridges the gap between the metric definition and the differential definition.

3. Inductive Approximation Scheme:

   • The main proofs proceed by induction on the dimension of the definable set.
   • They construct the approximation $g$ by gluing local approximations $g_S$ (defined via retractions $\pi_S$ onto strata) using the sharp partition of unity.
   • The error in the derivative is controlled by the sum of the error in the local approximation and the error introduced by the partition of unity's gradient.

3. Key Contributions and Results

Main Theorems

• Theorem 4.2 (Sharp Partitions of Unity): For any locally closed definable set $A$ and a $C^2$ stratification $\mathcal{S}$, there exists a definable $C^1$ partition of unity subordinate to a covering such that the gradient of each function is bounded by $\mu/d(x, S)$. This is a significant improvement over previous results (e.g., K-CPV) where the constant was fixed; here, the constant can be made arbitrarily small ($\mu$).

• Theorem 4.3 (Inner Lipschitz Approximation): Let $f: A \to \mathbb{R}^k$ be an inner Lipschitz definable mapping. For any $\epsilon > 0$ and $\mu > 0$, there exists a definable $C^1$ inner Lipschitz mapping $g$ such that:

  1. $|f - g| < \epsilon$ (uniform approximation).
  2. $|dg| \leq \mathring{L}f(x) + \mu$ (derivative bound).
  • This result holds for any polynomially bounded o-minimal structure.

• Corollary 4.4 ($C^\infty$ Approximation): If the o-minimal structure admits $C^\infty$ cell decomposition (e.g., semi-algebraic or globally subanalytic sets), the approximation $g$ can be chosen to be $C^\infty$.

• Theorem 4.6 & Corollary 4.9 (Outer Lipschitz Extension):

  • The results are extended to outer Lipschitz mappings (standard Euclidean Lipschitz).
  • If the structure admits $C^\infty$ cell decomposition, any definable Lipschitz map $f: A \to \mathbb{R}^k$ can be approximated by a $C^\infty$ definable Lipschitz map $g$ with $Lip(g) \leq Lip(f) + \mu$.
  • This relies on the definable version of Kirszbraun's Theorem (Theorem 4.7) to extend the domain to $\mathbb{R}^n$ before applying the approximation.

Specific Applications

• Weight Functions: The ability to define approximations with bounded derivatives is essential for defining weight functions used in studying solutions to Partial Differential Equations (PDEs) on singular domains.
• Mostowski's Separation Lemma: The authors provide a $C^\infty$ definable variant of Mostowski's separation lemma (Corollary 4.11), showing that disjoint closed definable sets in a manifold can be separated by a $C^\infty$ Lipschitz function.

4. Significance

1. Bridging Singularity and Smoothness: The paper successfully demonstrates that despite the rigidity of o-minimal structures (which often prevents $C^\infty$ partitions of unity), one can still achieve high-order smoothness ($C^\infty$) for approximations if the structure admits $C^\infty$ cell decomposition, provided one carefully controls the derivative bounds.
2. Inner Metric Analysis: By focusing on the inner metric, the authors address the specific geometry of singular spaces where the Euclidean metric fails to capture the connectivity of the space. This is vital for analysis on singular varieties.
3. Sharp Bounds: The construction of partitions of unity with arbitrarily small derivative constants (relative to the distance to the stratum) is a novel technical achievement. This allows for the preservation of Lipschitz constants during approximation, a property that was previously difficult to guarantee in the definable category.
4. Generalization: The results generalize previous work by Fisher (who handled outer Lipschitz in semi-algebraic settings) to the broader context of inner Lipschitz mappings in general polynomially bounded o-minimal structures.

In summary, this paper provides a robust toolkit for performing smooth analysis on singular definable sets, ensuring that the geometric and metric properties (specifically Lipschitz constants) are preserved during the approximation process.