$\mathrm{PGL}(3)$-invariant integrable systems from factorisation of linear differential and difference operators

This paper presents a unified framework for constructing continuous and discrete $\mathrm{PGL}(3)$-invariant integrable systems by generalizing the Schwarzian derivative and cross-ratio through the factorization of third-order linear spectral problems, thereby establishing dualities, exact discretizations, and a generating Lagrangian structure for the Boussinesq hierarchy.

The Gist

Imagine you are trying to describe the shape of a winding road. You could measure every single curve, hill, and dip with a ruler, but that data would be messy and specific to your exact starting point. Instead, a better way is to describe the road's intrinsic shape—how it bends relative to itself, regardless of where you stand or how you rotate your map.

This paper is about finding the "perfect map" for a very complex class of mathematical systems called integrable systems. These are equations that describe waves, fluids, and other physical phenomena that behave in a very orderly, predictable way (they don't just turn into chaos).

Here is the breakdown of what the authors did, using simple analogies:

1. The Problem: Too Many Perspectives

In mathematics, there are different "ranks" of complexity.

• Rank 2 (The KdV Equation): Think of this as a simple, one-lane road. Mathematicians have known for a long time how to describe this road using a special tool called the Schwarzian derivative. It's like a "shape-shifter" that stays the same even if you stretch or rotate your view of the road. This is the "Rank 2" world.
• Rank 3 (The Boussinesq or BSQ Equation): Now, imagine a three-lane highway where the lanes interact with each other. This is much more complex. While mathematicians have equations for this, they haven't had a unified, "shape-preserving" way to describe it that works for both continuous waves (like water) and discrete steps (like pixels on a screen).

The authors wanted to build a universal translator for this Rank 3 world, similar to the one they have for Rank 2.

2. The Tool: The "Magic Lens" (Spectral Problems)

To solve this, the authors used a concept called a linear spectral problem.

• The Analogy: Imagine you have a mysterious black box (the complex system). You can't see inside, but you can shine a light through it (the "spectral problem"). The way the light bends tells you about the box's internal structure.
• The Breakthrough: The authors realized that if you shine a "third-order" light (a more complex beam) through the Rank 3 system, you can extract specific numbers that never change, no matter how you twist or turn your coordinate system. These are the PGL(3)-invariants.

They created a new "magic lens" that generalizes the old Rank 2 tool. Instead of just one number describing the curve, they found two numbers ($z_1$ and $z_2$) that act as the fundamental coordinates for this 3D highway.

3. The Bridge: Folding the Paper (Factorization)

One of the coolest parts of the paper is how they connected the continuous world (smooth waves) with the discrete world (step-by-step grids).

• The Analogy: Imagine a piece of paper with a drawing on it. If you fold the paper perfectly along a line, the drawing on one side matches the drawing on the other side.
• The Math: The authors used a technique called factorization. They showed that the complex equations for the smooth waves can be "folded" into equations for the discrete steps, and vice versa. This revealed a duality: the continuous system and the discrete system are actually two sides of the same coin. If you understand one, you automatically understand the other.

4. The Result: A New Language for Waves

By using these new "invariant" coordinates ($z_1, z_2$), the authors wrote down new equations for the Rank 3 systems.

• The "Schwarzian" Upgrade: Just as the old Rank 2 equations had a "Schwarzian" version (a projective form), these new Rank 3 equations are the Rank 3 Schwarzian versions.
• The "Generating" System: They didn't just write one equation; they wrote a "Master Equation." Think of this as a Swiss Army Knife. By adjusting a few knobs (parameters) on this Master Equation, you can generate the entire family of Rank 3 wave equations, from the simplest to the most complex.

5. Why Does This Matter?

• Unification: It brings together geometry (shapes), algebra (equations), and physics (waves) under one roof.
• Predictability: Because these systems are "integrable," they are rare and special. Finding a unified way to describe them helps us understand how complex systems (like water waves or even light in certain materials) behave without turning into chaos.
• Future Proofing: The authors showed that their method isn't just for Rank 3. It's a recipe that can be used for Rank 4, Rank 5, and beyond. They built the ladder; now others can climb it.

Summary in a Nutshell

The authors took a messy, complex 3-lane highway of mathematical equations and found a special set of coordinates that make the road look the same from any angle. They discovered that the smooth version of this road and the pixelated version are actually the same thing, just viewed differently. Finally, they built a "Master Key" (a generating system) that unlocks the entire family of these complex wave equations, providing a unified language for mathematicians and physicists to describe the universe's most orderly waves.

Technical Summary

1. Problem Statement

The paper addresses the lack of a systematic projective formulation for rank-3 integrable systems, specifically the Boussinesq (BSQ) hierarchy.

• Context: While the rank-2 case (KdV hierarchy) is well-understood through the lens of $PGL(2)$-invariance (e.g., the Schwarzian derivative and the Schwarzian KdV equation), the rank-3 case ($PGL(3)$) has remained underdeveloped.
• Gap: Existing formulations of BSQ equations often rely on potentials ($u, v$) rather than projective invariants. Although $PGL(2)$-invariant "Schwarzian BSQ" equations exist, a unified $PGL(3)$-invariant framework that naturally generalizes the Schwarzian KdV and cross-ratio equations to the rank-3 setting was missing.
• Goal: To construct a unified approach for both continuous and discrete $PGL(3)$-invariant integrable systems using linear spectral problems and their factorization, providing explicit formulas for invariants and deriving the associated BSQ systems.

2. Methodology

The authors employ a unified algebraic and geometric approach based on linear spectral problems and their factorization.

• Spectral Problems: The study starts with third-order linear differential operators (continuous) and difference operators (discrete).
  • Continuous: $(\partial_x^3 + u\partial_x + v)\phi = \lambda\phi$
  • Discrete: $(T^3 + hT^2 + gT + \alpha)\phi = \lambda\phi$
• Projective Coordinates: The authors introduce inhomogeneous coordinates $z_1 = \phi_1/\phi_3$ and $z_2 = \phi_2/\phi_3$ derived from the linearly independent solutions of the spectral problems. These coordinates serve as the fundamental dependent variables.
• Factorization and Duality: A core methodological tool is the factorization of the spectral operators.
  • Factorizing the continuous operator induces a Darboux transformation, which maps the continuous problem to a discrete one (exact discretization).
  • Factorizing the discrete operator in a second direction induces a self-duality, revealing the multi-dimensional consistency (3D-consistency) of the lattice systems.
• Invariant Theory: The authors utilize Lie group theory (specifically $PGL(3)$ actions) to derive generating differential and difference invariants. They generalize the Schwarzian derivative (rank-2) and the cross-ratio (discrete rank-2) to the rank-3 setting.

3. Key Contributions and Results

A. Construction of $PGL(3)$ Invariants

The paper provides explicit formulae for the generating sets of invariants:

• Differential Invariants (Continuous): Two invariants, $S_1[z_1, z_2]$ and $S_2[z_1, z_2]$, are derived.
  • $S_1$ generalizes the Schwarzian derivative (weight 2).
  • $S_2$ is a weight-3 invariant related to the projective arc-length.
  • They are expressed as rational functions of derivatives of $z_1, z_2$ involving determinants (e.g., $|z^{(4)}, z^{(1)}| / |z^{(2)}, z^{(1)}|$).
  • Composition Rules: The paper derives the composition rules (connection formulas) for these invariants under reparametrization $x \to f(x)$, generalizing the classical Schwarzian composition rule.
• Difference Invariants (Discrete): Two invariants, $I_1[z_1, z_2]$ and $I_2[z_1, z_2]$, are constructed using determinants of shifted vectors. These generalize the cross-ratio.
  • The paper establishes the continuum limits of these difference invariants, showing they recover the differential invariants $S_1$ and $S_2$ as the lattice spacing $\delta \to 0$.

B. Derivation of $PGL(3)$-Invariant BSQ Systems

Using the compatibility conditions of the Lax pairs, the authors derive the integrable systems in terms of $z_1, z_2$:

• Continuous System: A coupled system of PDEs (Eq. 4.2) governing the evolution of $z_1, z_2$. This system is manifestly $PGL(3)$-invariant and represents the rank-3 analogue of the Schwarzian KdV equation.
  • Decoupling Mechanism: A "lifting-decoupling" mechanism is introduced, showing that each component ($z_1$ or $z_2$) independently satisfies the Schwarzian BSQ equation (a $PGL(2)$-invariant equation). This explains the reduction from rank-3 to rank-2.
• Discrete System: A coupled system on a 5-point stencil (Eq. 4.13) defined on a square lattice.
  • This system is shown to be the exact discretization of the continuous system.
  • It is lifted to a three-component quad-system (Eq. 4.21) which is multi-dimensionally consistent (3D-consistent) on a cube. This system serves as the rank-3 analogue of the cross-ratio equation (Q1 in the ABS classification).
  • Like the continuous case, individual components satisfy the discrete Schwarzian BSQ equation.

C. Generating Systems and Lagrangian Structure

The paper derives generating equations that encode the entire BSQ hierarchy:

• Semi-discrete Generating System: A non-autonomous differential-difference system (Eq. 4.38) derived by deforming the spectral problem with respect to lattice parameters.
• Generating PDE: A coupled system of fourth-order PDEs (Eq. 4.67) where the lattice parameters serve as independent variables ($s, t$).
  • This system is derived from the compatibility of deformed spectral problems.
  • Lagrangian Structure: The authors present a Lagrangian (Eq. 4.68) for this generating PDE. They prove that the Lagrangian is $PGL(3)$-invariant up to null Lagrangian terms (divergence terms), ensuring the Euler-Lagrange equations are invariant.
  • This generating PDE is linked to the Einstein–Maxwell–Weyl theory, suggesting deep physical connections between water wave theory (BSQ) and gravitational waves.

4. Significance and Implications

• Unified Framework: The paper successfully unifies the continuous and discrete theories of rank-3 integrable systems under the $PGL(3)$ symmetry group, filling a significant gap in the literature between the well-known rank-2 (KdV) and higher-rank theories.
• Geometric Insight: It provides a clear geometric interpretation where the BSQ flow governs the evolution of non-degenerate projective curves in $\mathbb{P}^2$. The "lifting-decoupling" mechanism offers a new perspective on how higher-rank systems reduce to lower-rank ones.
• Multi-dimensional Consistency: The identification of the discrete system as a 3D-consistent quad-system places the BSQ hierarchy firmly within the modern theory of discrete integrable systems (ABS classification), offering a natural rank-3 extension of the cross-ratio equation.
• Future Directions: The framework is generalizable to arbitrary rank $N$ ($PGL(N)$), offering a systematic path to constructing $PGL(N)$-invariant integrable systems. It also opens avenues for studying higher-rank Painlevé equations and connections to matrix Schwarzian derivatives and Grassmannian curves.

In summary, this work establishes a rigorous algebraic and geometric foundation for $PGL(3)$-invariant integrable systems, providing explicit invariants, integrable equations, and generating hierarchies that generalize the classical Schwarzian theory to the rank-3 Boussinesq setting.