Spectral Properties of Off Diagonal Block Linear Relations via Moore Penrose Inverses in Hilbert Spaces

This paper characterizes the essential spectra and resolvent sets of off-diagonal block linear relations in Hilbert spaces by establishing precise spectral relationships with the products $AB$ and $BA$, and extends these results to relations involving Moore–Penrose inverses of closed linear operators with closed ranges.

The Gist

Imagine you are a detective trying to solve a mystery about a complex machine. This machine isn't made of gears and wires, but of abstract mathematical rules called Linear Relations. These rules describe how inputs (like numbers or vectors) transform into outputs. Sometimes, these rules are messy, broken, or even have multiple possible answers for a single input.

This paper is about understanding the "health" and "behavior" of a specific type of machine: a Block Linear Relation. Think of this machine as a two-story building with a special layout:

[ 0   A ]
[ B   0 ]

The top-left and bottom-right corners are empty (zeros). The action happens only on the off-diagonal: A sends things from the second floor to the first, and B sends things from the first floor to the second.

Here is the breakdown of the paper's discoveries, explained with everyday analogies:

1. The "Moore-Penrose Inverse": The Universal Fix-It Tool

In the world of math, sometimes a machine breaks. It might be jammed (not invertible) or have too many answers (not unique). The Moore-Penrose Inverse (denoted as $T^\dagger$) is like a "best-effort" repair kit.

• The Analogy: Imagine you have a blurry photo (the original relation). You can't perfectly restore it to the original sharp image, but the Moore-Penrose Inverse gives you the sharpest possible version that makes the most sense mathematically. It filters out the noise and gives you the cleanest answer.
• The Paper's Contribution: The author, Arup Majumdar, figured out exactly how to build this "repair kit" for these complex, multi-valued machines (linear relations) in a Hilbert space (a fancy, infinite-dimensional playground for math). He proved that if you have a closed, well-behaved machine, you can always find this "best-fit" inverse, and he mapped out exactly how it behaves.

2. The "Echo Chamber" Effect (Spectral Properties)

The main mystery the paper solves is: If we know how the "A-to-B" loop and the "B-to-A" loop behave, can we predict how the whole building behaves?

• The Analogy: Imagine you are in a hallway with two mirrors facing each other (A and B).
  • If you shout a sound (a number $\lambda$) into the hallway, it bounces off A, then B, then A again.
  • The paper asks: "If I know the 'echo patterns' (spectra) of the sound bouncing between A and B separately, can I predict the echo patterns of the whole hallway?"
• The Discovery: Yes! The paper proves a precise rule. The "echo patterns" (specifically the Resolvent Set and Essential Spectra) of the whole building depend entirely on the square of the echo patterns of the individual loops ($AB$ and $BA$).
  • If the loop $A \to B$ is stable, and $B \to A$ is stable, then the whole building is stable.
  • If the loops have a "glitch" (a specific type of instability), the whole building will have a glitch at the square root of that frequency.

3. The Special Case: The "Self-Correcting" Machine

The paper gets even more interesting when it looks at a special machine where one part is the original relation ($T$) and the other part is its "repair kit" ($T^\dagger$).

• The Analogy: Imagine a machine $T$ that takes a messy input and tries to clean it. The Moore-Penrose Inverse $T^\dagger$ is the tool that tries to reverse that cleaning process.
• The Setup: The paper builds a machine where the top floor is the "Cleaner" ($T^\dagger$) and the bottom floor is the "Mess-maker" ($T$).
• The Result: By using the rules established earlier, the author can now predict exactly when this "Self-Correcting" machine will work perfectly and when it will fail, just by looking at how $T$ and its repair kit interact.

4. Why This Matters (The "So What?")

You might wonder, "Why do we care about these abstract buildings?"

• Real-World Connection: These mathematical structures are used in Digital Image Restoration (fixing blurry photos), Linear Systems Theory (controlling robots or satellites), and Regression Analysis (predicting stock markets or weather).
• The Takeaway: When engineers or scientists are trying to fix a broken system or predict its behavior, they often deal with systems that aren't perfectly reversible. This paper gives them a new, powerful map. It tells them: "Don't look at the whole giant, confusing machine. Just look at the two smaller loops inside it. If you understand those, you understand the whole thing."

Summary in a Nutshell

The paper is a guidebook for understanding complex, two-way mathematical machines. It introduces a "best-fit" repair tool (the Moore-Penrose Inverse) for broken machines and proves that the behavior of a two-story machine is simply the square of the behavior of the loops connecting its floors. This allows mathematicians and scientists to solve huge, complicated problems by breaking them down into smaller, manageable pieces.

Technical Summary

1. Problem Statement

The paper addresses two interconnected problems in the spectral theory of linear relations within Hilbert spaces:

1. Spectral Characterization of Block Operators: Determining the essential spectra and resolvent sets of off-diagonal block linear relations of the form $T = \begin{bmatrix} 0 & A \\ B & 0 \end{bmatrix}$ in terms of the spectral properties of the compositions $AB$ and $BA$.
2. Moore–Penrose Inverses of Linear Relations: Extending the theory of Moore–Penrose inverses to closed linear relations (which generalize linear operators and may have non-dense domains or multi-valued parts) and applying these inverses to analyze specific block structures where one entry is the Moore–Penrose inverse of the other ($T^\dagger$).

The study is motivated by the need to handle operators that may not be densely defined or invertible in the traditional sense, which is common in applications involving differential equations and singular systems.

2. Methodology

The author employs a rigorous functional analytic approach combining:

• Linear Relation Theory: Utilizing the definitions of domains ($D(T)$), ranges ($R(T)$), kernels ($N(T)$), and multi-valued parts ($M(T)$) for linear relations. The paper distinguishes between closed linear relations ($CR(H, K)$) and bounded/continuous linear relations.
• Moore–Penrose Generalization: Defining the Moore–Penrose inverse $T^\dagger$ for a closed linear relation $T$ as $T^\dagger = P_{N(T)^\perp} T^{-1} P_{R(T)}$, where $P$ denotes orthogonal projections. The paper investigates the algebraic and topological properties of this inverse.
• Spectral Decomposition: Analyzing the resolvent set $\rho(T)$ and various essential spectra ($\sigma_{e1}, \sigma_{e2}, \sigma_{e3}, \sigma_{e4}$) defined via semi-Fredholm properties ($\Phi_+$ and $\Phi_-$).
• Block Matrix Algebra: Decomposing the square of the block relation $T^2$ to relate it to the diagonal block matrix $\begin{bmatrix} AB & 0 \\ 0 & BA \end{bmatrix}$, allowing the transfer of spectral properties from the product relations to the block relation.

3. Key Contributions and Results

A. Characterization of Moore–Penrose Inverses for Closed Linear Relations

The paper establishes fundamental properties of $T^\dagger$ for $T \in CR(H, K)$:

• Adjoint Property: Proves that $(T^\dagger)^* = (T^*)^\dagger$ when $T$ has a closed range.
• Boundedness: Shows that if $R(T)$ is closed, $T^\dagger$ is a bounded operator.
• Algebraic Identities: Establishes generalized Penrose equations, such as $TT^\dagger T = T$ and $T^\dagger TT^\dagger = T^\dagger$.
• Product Rules: Derives conditions under which $(TS)^\dagger = S^\dagger T^\dagger$ (specifically when $R(S) = N(T)^\perp$) and provides inclusion relations for $(T^*T)^\dagger$.
• Direct Sums: Proves that the Moore–Penrose inverse commutes with direct sums: $(T_1 \oplus T_2)^\dagger = T_1^\dagger \oplus T_2^\dagger$.
• Reduced Minimal Modulus: Relates the reduced minimal modulus $\gamma(T)$ to the norm of the inverse: $\gamma(T) = 1/\|T^\dagger\|$ for closed range relations.

B. Spectral Properties of Off-Diagonal Block Relations

The paper characterizes the spectra of $T = \begin{bmatrix} 0 & A \\ B & 0 \end{bmatrix}$ where $A \in CR(K, H)$ and $B \in CR(H, K)$:

• Resolvent Set: The resolvent set is characterized as:
  $$\rho(T) = \{ \lambda \in \mathbb{C} : \lambda^2 \in \rho(AB) \cap \rho(BA) \}$$
  This result holds provided $AB$ and $BA$ are closed linear relations.
• Essential Spectra: The paper derives precise formulas for the essential spectra ($\sigma_{e1}, \sigma_{e2}, \sigma_{e3}, \sigma_{e4}$) in terms of the essential spectra of the products $AB$ and $BA$. For example:
  $$\sigma_{e1}(T) = \{ \lambda \in \mathbb{C} : \lambda^2 \in \sigma_{e1}(AB) \cup \sigma_{e1}(BA) \}$$
  Similar union-based characterizations are provided for $\sigma_{e2}$ and $\sigma_{e4}$, while $\sigma_{e3}$ involves intersections and unions of the component spectra.

C. Application to $T^\dagger$-Based Block Relations

A significant application involves the specific block relation $A = \begin{bmatrix} 0 & T^\dagger \\ T & 0 \end{bmatrix}$, where $T$ is a closed, continuous linear relation with a closed range.

• The paper applies the general spectral results to this specific case, characterizing $\rho(A)$ and $\sigma_{ei}(A)$ in terms of the spectral properties of the products $TT^\dagger$ and $T^\dagger T$.
• It is shown that $TT^\dagger$ and $T^\dagger T$ are bounded operators (projections), simplifying the spectral analysis of the block matrix.

4. Significance and Impact

• Generalization: The work extends classical results on block operator matrices (typically defined for bounded operators with dense domains) to the broader class of closed linear relations. This is crucial for handling operators with non-dense domains or multi-valued parts, which arise in boundary value problems and control theory.
• Unified Framework: By linking the spectral properties of block structures to the compositions of their entries, the paper provides a unified framework for analyzing complex structured systems.
• Moore–Penrose Theory: The rigorous treatment of Moore–Penrose inverses for linear relations fills a gap in the literature, providing necessary algebraic tools (like the product rule and adjoint property) that were previously less developed for this class of objects.
• Foundation for Future Research: The results on the reduced minimal modulus and direct sums of inverses provide a foundation for further investigations into the stability and perturbation of structured linear relations in Hilbert spaces.

In summary, Majumdar's paper successfully bridges the gap between the algebraic theory of generalized inverses and the spectral theory of block linear relations, offering precise characterizations that are applicable to a wider class of operators than previously possible.