Raja's covering index of $L_p$ spaces

This paper computes Raja's covering index for infinite-dimensional Hilbert spaces, establishes sharp asymptotic estimates for scalar-valued $L_p$ spaces under renormability conditions, provides uniform upper bounds for Bochner spaces that partially resolve a question by Raja, and derives power-type lower bounds for non-commutative $L_p$ spaces using non-commutative Clarkson inequalities.

The Gist

Imagine you have a giant, perfect sphere (like a beach ball) made of a special, stretchy material. This sphere represents the "unit ball" of a mathematical space called a Banach space. Mathematicians love to study these shapes because they hold the secrets to how different kinds of infinite-dimensional spaces behave.

Now, imagine you want to cut this giant beach ball into $n$ pieces using flat, smooth slices (convex sets). You want to do this in a way that every single piece is as "round" and "bulky" as possible in the middle.

The paper you're asking about is about a new ruler invented by a mathematician named Raja. This ruler, called the Covering Index ($\Theta_X(n)$), measures exactly how "bulky" the biggest round part inside your cut pieces can be.

• High Index: Your pieces are still very round and full of space. The ball is hard to break down.
• Low Index: Your pieces are flat, thin, or weirdly shaped. The ball is easy to slice up.

The authors, Tomasz Kania and Natalia Maślany, used this ruler to measure three different types of mathematical "beach balls" to see how they react when sliced.

1. The Perfect Sphere: Hilbert Spaces

First, they looked at Hilbert spaces. Think of this as the most perfect, symmetrical sphere you can imagine (like the space $\ell_2$ or standard 3D space, but with infinite dimensions).

• The Question: If you slice this perfect sphere into 2 pieces, how round can the biggest round part inside those pieces be? Raja asked for the exact number.
• The Discovery: The authors found the exact answer! If you slice it into $n$ pieces, the "roundness" drops exactly by the square root of $n$.
  • Analogy: Imagine a perfectly round cake. If you cut it into 4 pieces, the biggest round cookie you can cut out of a slice is half the size of the original cake. If you cut it into 9 pieces, it's one-third the size.
  • The Result: For 2 pieces, the roundness is $1/\sqrt{2}$ (about 0.707). This solves a specific puzzle Raja had.

2. The Stretchy Rubber Sheet: Scalar $L_p$ Spaces

Next, they looked at $L_p$ spaces. These are like the Hilbert sphere, but the material is different.

• If $p=2$, it's the perfect sphere.
• If $p=1$, it's a diamond shape (very pointy).
• If $p$ is huge, it looks like a cube (very flat sides).

The authors asked: "If we slice these different shapes, how does the roundness change?"

• The Discovery: They built a specific recipe for slicing these shapes. They found that the roundness drops based on the power of $p$.
  • Analogy: Imagine a rubber sheet. If you stretch it out (changing $p$), it becomes harder to keep a round shape inside a slice. The "roundness" drops by the $p$-th root of $n$.
  • The Result: For these spaces, the index is roughly $1/n^{1/p}$. This is the best possible rate; you can't do better than this.

3. The "Universal" Wrapper: Bochner Spaces

This is the most surprising part. Imagine taking your rubber sheet ($L_p$) and wrapping it around a weird, lumpy object (another space called $E$). This creates a Bochner space ($L_p(\mu; E)$).

• The Question: Does the weird shape of the object inside ($E$) change how easy it is to slice the wrapper? Raja wondered if the "inner geometry" of the object mattered.
• The Discovery: No, it doesn't matter!
  • Analogy: Imagine wrapping a gift. Whether you wrap a perfect sphere, a jagged rock, or a squishy pillow inside the paper, the paper itself behaves the same way when you try to cut it. The "roundness" of the cut pieces depends only on the paper (the $L_p$ part), not the object inside.
  • The Result: Even if the inner object is very strange and hard to smooth out, the covering index still drops at the same rate ($1/n^{1/p}$). This answers Raja's question with a "No": the covering index doesn't always tell you everything about the inner shape.

4. The Quantum Sphere: Non-Commutative Spaces

Finally, they looked at Non-Commutative $L_p$ spaces. These are like "quantum" versions of the shapes above, where the order of operations matters (like in quantum physics).

• The Discovery: They couldn't find the exact answer for these weird quantum shapes, but they found a safety net. They proved that no matter what, the roundness cannot drop faster than a certain speed ($1/n^{1/r}$).
• The Result: It's like saying, "We don't know the exact speed of this quantum car, but we know it can't go slower than 50 mph."

Summary: Why Does This Matter?

Mathematicians use these "covering indices" to understand the hidden structure of infinite-dimensional spaces.

• Hilbert spaces are the most "honest" and predictable.
• $L_p$ spaces follow a strict rule based on their "stretchiness" ($p$).
• Bochner spaces teach us that sometimes, the outer layer hides the inner complexity completely.
• Non-commutative spaces remind us that in the quantum world, things are still a bit of a mystery, but we have lower limits.

In short, this paper gives us a precise ruler to measure how "chunky" or "flat" mathematical spaces are when we try to break them apart, revealing that some shapes are much more resistant to being sliced than others.

Technical Summary

Here is a detailed technical summary of the paper "Raja's Covering Index of $L^p$ Spaces" by Tomasz Kania and Natalia Maślany.

1. Problem Statement and Context

The paper investigates Raja's covering index, denoted as $\Theta_X(n)$, a quantitative invariant for Banach spaces introduced in a recent preprint by Raja [5].

• Definition: For a Banach space $X$, $\Theta_X(n)$ is the infimum of the maximum "essential inradius" over all coverings of the unit ball $B_X$ by $n$ closed convex sets.
• Essential Inradius ($\varrho(A)$): The supremum of radii $r$ such that a set $A$ contains a ball of radius $r$ centered at some point, restricted to an infinite-dimensional subspace of finite codimension.
• Significance: The covering index measures the resistance of the unit ball to decomposition into convex parts. It is closely linked to asymptotic uniform smoothness (AUS) and the Szlenk index.
• Open Questions: Raja established that if a space admits an equivalent $p$-AUS norm, then $\Theta_X(n) \gtrsim n^{-1/p}$. However, exact values and sharp constants for specific classical spaces (like Hilbert spaces and $L^p$ spaces) were unknown. Specifically, Raja asked for the precise value of $\Theta_{\ell_2}(2)$ and questioned the extent to which the covering index is governed by the asymptotic moduli of convexity and smoothness.

2. Methodology

The authors employ a combination of geometric analysis, functional analytic constructions, and specific inequalities:

1. Goal Derivation (Szlenk Index): For the lower bounds in Hilbert spaces, they utilize Raja's "goal derivation" technique, explicitly computing the iterative removal of points from the unit ball based on the essential inradius.
2. Block Decomposition: For $L^p$ spaces, the core methodology involves constructing explicit coverings of the unit ball using orthogonal (or disjoint) decompositions of the underlying measure space. By partitioning the domain $\Omega$ into $n$ disjoint sets, they define subsets of the unit ball where the norm on each component is bounded.
3. Non-Commutative Inequalities: For non-commutative $L^p$ spaces, they rely on non-commutative Clarkson inequalities to establish uniform smoothness properties, which are then fed into Raja's general lower bound theorem.
4. Vector-Valued Extension: They extend the scalar block decomposition to Bochner spaces $L^p(\mu; E)$, demonstrating that the upper bound depends only on $p$ and not on the geometry of the Banach space $E$.

3. Key Contributions and Results

A. Exact Calculation for Hilbert Spaces

The authors provide the first exact computation of the covering index for infinite-dimensional Hilbert spaces ($H$).

• Result: For every $n \in \mathbb{N}$,
  $$\Theta_H(n) = n^{-1/2}$$
• Specific Answer: This resolves Raja's Problem 4.6, proving that $\Theta_{\ell_2}(2) = 1/\sqrt{2}$.
• Proof Strategy:
  • Upper Bound: Constructed via an orthogonal decomposition of $H$ into $n$ infinite-dimensional subspaces.
  • Lower Bound: Derived by explicitly computing the goal derivation on Hilbert balls, showing that the $n$-th iteration of the derivation is non-empty if $\epsilon < n^{-1/2}$.

B. Sharp Asymptotics for Scalar $L^p(\mu)$ Spaces

For scalar-valued Lebesgue spaces $L^p(\mu)$ ($1 \le p < \infty$), the authors establish sharp asymptotic behavior.

• Upper Bound: They construct an explicit block decomposition yielding:
  $$\Theta_{L^p(\mu)}(n) \le n^{-1/p}$$
  This holds for all $n \in \mathbb{N}$ and is independent of the measure space structure (provided it is infinite-dimensional).
• Lower Bound: Under the hypothesis that $L^p(\mu)$ admits an equivalent $p$-AUS norm (which is true for $1 < p < \infty$in standard cases like$\ell_p$and$L^p[0,1]$), Raja's general theorem provides:
  $$\Theta_{L^p(\mu)}(n) \gtrsim n^{-1/p}$$
• Conclusion: For $1 < p < \infty$, the covering index decays exactly as:
  $$\Theta_{L^p(\mu)}(n) \asymp n^{-1/p}$$

C. Vector-Valued Bochner Spaces and Raja's Problem 4.7

The authors address Problem 4.7 from Raja's work, which asks if the covering index is strictly governed by the asymptotic moduli of uniform convexity and smoothness.

• Result: They prove that for any non-zero Banach space $E$ and non-atomic $\sigma$-finite measure space $(\Omega, \Sigma, \mu)$:
  $$\Theta_{L^p(\mu; E)}(n) \le n^{-1/p}$$
• Significance: This bound is independent of the geometry of $E$.
  • They construct examples where $E_1$ is $p$-AUS (e.g., $E_1 = \mathbb{K}$) and $E_2$ is not AUSable (e.g., a space with Szlenk index $>\omega$).
  • Despite the drastic difference in the asymptotic geometry of $E_1$ and $E_2$, the covering index of $L^p(\mu; E_1)$ and $L^p(\mu; E_2)$ shares the same power-type upper decay rate ($n^{-1/p}$).
• Implication: This provides a partial negative answer to Raja's Problem 4.7, demonstrating that the covering index's upper bounds are not solely determined by the asymptotic moduli of the space.

D. Non-Commutative $L^p$ Spaces

The paper extends the analysis to non-commutative $L^p$ spaces $L^p(M, \tau)$ associated with semifinite von Neumann algebras.

• Result: Using non-commutative Clarkson inequalities, they show these spaces are $r$-AUS with $r = \min\{p, 2\}$. Consequently:
  $$\Theta_{L^p(M, \tau)}(n) \gtrsim n^{-1/r}$$
• Limitation: The authors note that this lower bound is not sharp when $p > 2$ (where $r=2$ but the optimal decay is likely $n^{-1/p}$), and they do not attempt to optimize the exponent or constants in the non-commutative setting.

4. Significance

1. Resolution of Open Problems: The paper definitively answers Raja's question regarding the exact value of $\Theta_{\ell_2}(2)$ and establishes the precise decay rate for classical $L^p$ spaces.
2. Geometric Insight: It reveals a decoupling between the asymptotic smoothness of a Banach space $E$ and the covering index of the Bochner space $L^p(\mu; E)$. The covering index is shown to be robust against changes in the asymptotic geometry of the vector-valued component, provided the measure space is non-atomic.
3. Methodological Advancement: The explicit construction of block decompositions for the unit ball provides a concrete tool for estimating covering indices, moving beyond abstract existence proofs.
4. Foundation for Future Work: The paper highlights the gap between known lower bounds (via AUS) and upper bounds in non-commutative settings, identifying the determination of sharp exponents for Schatten classes and other non-commutative spaces as a significant open problem.

In summary, Kania and Maślany successfully map the covering index landscape for $L^p$ spaces, proving that while the index is tightly linked to the exponent $p$, it is surprisingly insensitive to the specific asymptotic geometry of the underlying vector space in the Bochner setting.