On the role of the slowest observable in one-dimensional Markov processes to construct quasi-exactly-solvable generators with $N=2$ explicit levels

This paper demonstrates that constructing quasi-exactly-solvable generators with two explicit energy levels is more intuitive and technically simpler when approached through one-dimensional Markov processes by treating the slowest observable as the central object from which all other properties can be reconstructed.

The Gist

Imagine you are watching a crowded room of people (the Markov process) trying to find their way to a specific, comfortable spot (the steady state).

In physics, we usually try to predict exactly where everyone will be at any given moment. This is like solving a massive, complex puzzle. Sometimes, the puzzle is so hard we can only solve a tiny piece of it. This is called a "Quasi-Exactly-Solvable" model: we can't see the whole picture, but we can clearly see the first two pieces.

This paper, by Cécile Monthus, offers a new, simpler way to look at these puzzles. Instead of staring at the people (the probabilities), the author suggests we focus on the slowest thing that changes in the room.

Here is the breakdown using everyday analogies:

1. The Two Key Characters

In any system trying to settle down, there are two main "actors" we care about:

• The Ground State (The Destination): This is the final, calm state where everyone is settled. In the paper, this is called $E_0 = 0$. It's like the room being perfectly quiet.
• The Slowest Observable (The "Relaxation Meter"): This is the second most important thing. It tells us how fast the room is getting quiet. If you shout in the room, this is the echo that takes the longest to fade away. The paper calls this $L_1(x)$.

The Big Idea: The author says, "Stop trying to solve the whole room at once. Just focus on this 'Relaxation Meter' ($L_1$). If you know how this one thing behaves, you can actually rebuild the entire puzzle for the first two levels of the system."

2. The Two Perspectives: Quantum vs. Markov

The paper bridges two worlds:

• The Quantum World (The Map): Physicists often use "Quantum Mechanics" to describe these systems. They look at a map with hills and valleys (potentials) and try to find the exact paths (eigenstates) a particle takes. It's like trying to draw every single possible route a hiker could take on a mountain.
• The Markov World (The Flow): This paper looks at the system as a flow of probability, like water flowing through pipes or people moving through a hallway. It focuses on Currents (how much stuff is moving) and Divergence (where stuff is piling up or leaving).

The Analogy:
Imagine a river.

• The Quantum view asks: "What is the exact shape of the water surface at every single point?"
• The Markov view asks: "How fast is the water flowing, and where is it slowing down?"

The author argues that the "Flow" view is much more intuitive. It's easier to understand the river by watching the current than by measuring the surface tension of every drop.

3. The "Doob Transformation": The Magic Trick

The paper introduces a mathematical tool called a Doob Transformation. Think of this as a special pair of glasses or a filter.

• Before the glasses: You see a system where the "slowest echo" ($L_1$) is fading away. It's a bit messy.
• After the glasses: You apply the Doob transformation. Suddenly, the "slowest echo" becomes the new "ground state" (the new destination). The system looks different, but it's mathematically equivalent.

Why is this cool?
It's like taking a photo of a messy room, applying a filter that makes the mess look like a new, organized room, and realizing that the rules for the new room are much simpler to write down. This allows the author to construct new, solvable models just by picking a simple shape for the "slowest observable" and letting the math fill in the rest.

4. Changing the Lens (Variables)

The paper also discusses changing the "lens" through which we view the system.

• Variable $y$: Imagine stretching the room so that the "Relaxation Meter" becomes a straight line. Suddenly, the complex curves of the system become simple straight lines.
• Variable $z$: Imagine changing the floor so that the "friction" (diffusion) is the same everywhere. This makes the math look like standard textbook problems.

By switching to these specific lenses, the complex, scary equations turn into simple, familiar shapes (like polynomials or trigonometric waves).

Summary: What did we learn?

1. Focus on the Slowest: Instead of trying to solve the whole system, identify the "slowest observable" (the thing that takes the longest to settle).
2. Build from the Bottom Up: If you define this slowest thing, you can mathematically reconstruct the entire system's rules (the forces, the probabilities, the energy levels) for the first two states.
3. Flow is Easier than Maps: Looking at the system as a flow of currents (Markov) is often more intuitive and mathematically simpler than looking at it as a static map of energy levels (Quantum).
4. The Magic Filter: Using the Doob transformation allows us to turn a complex, unsolvable problem into a new, simpler problem that we can solve.

In a nutshell: The paper is a guide on how to stop trying to solve the whole puzzle at once. Instead, it teaches us to find the "key piece" (the slowest observable), use a special mathematical filter to rearrange the puzzle, and suddenly, the solution for the first two pieces becomes obvious.

Technical Summary

1. Problem Statement

The paper addresses the construction of Quasi-Exactly-Solvable (QES) models in one-dimensional quantum mechanics and statistical physics.

• Context: In QES systems, only a finite number $N$ of eigenstates can be computed explicitly. While the case $N=1$ (ground state only) is trivial, the case $N=2$ (ground state $\Phi_0$ and first excited state $\Phi_1$) is the simplest non-trivial case of significant interest.
• Challenge: Traditional constructions rely on supersymmetric quantum mechanics (SUSY QM), involving the factorization of Hamiltonians ($H = Q^\dagger Q$) and the recursive construction of partner potentials. This approach can be algebraically complex and lacks intuitive physical interpretation regarding the dynamics of the system.
• Goal: To revisit the construction of $N=2$ QES models from the perspective of one-dimensional Markov processes satisfying detailed balance. The aim is to demonstrate that treating the slowest observable (the ratio of the first excited state to the ground state) as the central object simplifies the construction and provides new physical insights.

2. Methodology

The author employs a unified framework that bridges Markov processes and quantum mechanics via similarity transformations.

A. Unified Notation

The paper introduces a unified operator notation $\nabla$ and two potentials, $U(x)$ and $U_I(x)$, applicable to both:

1. Continuous Space: Where $\nabla = \partial_x$ (Fokker-Planck dynamics).
2. Discrete Space: Where $\nabla$ is a finite-difference operator (Markov jump processes on a lattice).

The Markov generator $G$ governing probability evolution $\partial_t P_t = G P_t$ is factorized as:
$$G = -\nabla J = \nabla e^{-U_I(x)} \nabla e^{U(x)}$$
where $J$ is the current operator. This factorization mirrors the SUSY QM factorization $H = Q^\dagger Q$.

B. The Core Concept: The Slowest Observable

The central object of the study is the slowest observable $L_1(x)$, defined as the ratio of the first excited quantum eigenstate to the ground state:
$$L_1(x) = \frac{\Phi_1(x)}{\Phi_0(x)}$$
In the Markov context, $L_1(x)$ is the left eigenvector of the generator $G$ (specifically of the adjoint $G^\dagger$) associated with the non-zero eigenvalue $E_1$ (the relaxation rate).

• $E_0 = 0$ corresponds to the steady state $P^*(x)$ and probability conservation.
• $E_1 > 0$ governs the exponential relaxation to the steady state.

C. The Construction Strategy

Instead of starting with a potential and solving for eigenstates, the paper proposes an inverse construction:

1. Input: Choose the diffusion coefficient $D(x)$ (or jump rates), the energy gap $E_1$, and the functional form of the slowest observable $L_1(x)$.
2. Constraint: $L_1(x)$ must have a single root (node) and a strictly positive derivative (to ensure the partner ground state has no nodes).
3. Reconstruction: Derive the steady state $P^*(x)$, the force fields, and the full generator $G$ directly from $L_1(x)$ and $E_1$.

D. Doob Transformation and Partner Systems

The paper utilizes the Doob transformation to construct a new Markov generator $G^{[1]}$ from the "partner" generator $\tilde{G}$.

• The partner generator $\tilde{G} = -J\nabla$ governs the dynamics of currents.
• Using the left eigenvector $\tilde{L}_1(x) = \nabla L_1(x)$ of $\tilde{G}$, a new Markov generator $G^{[1]}$ is constructed via similarity transformation. This $G^{[1]}$ corresponds to the quantum Hamiltonian $H^{[1]} = \tilde{H} - E_1$, which has a vanishing ground state energy.

3. Key Contributions

1. Re-interpretation of SUSY QM via Markov Processes

The paper establishes a direct dictionary between SUSY QM concepts and Markov dynamics:

• Factorization: $H = Q^\dagger Q$ corresponds to the continuity equation $G = -\nabla J$.
• Partner Hamiltonian: The quantum partner $\tilde{H} = QQ^\dagger$ corresponds to the current dynamics generator $\tilde{G} = -J\nabla$.
• Recursion: The supersymmetric recursion $H^{[1]} = \tilde{H} - E_1$ is shown to be equivalent to the Doob transformation of the Markov generator, which produces a new physical process with a new steady state.

2. The "Slowest Observable" as the Primary Variable

The paper argues that choosing $L_1(x)$ is more intuitive than choosing the potential $V(x)$ or the force $F(x)$.

• Physical Meaning: $L_1(x)$ represents the observable that relaxes most slowly to equilibrium.
• Technical Simplicity: By fixing $L_1(x)$, the reconstruction of the steady state $P^*(x)$ becomes a direct integration problem (Eq. 97), avoiding the need to solve non-linear Riccati equations for the potential directly.

3. Generalization to Continuous and Discrete Spaces

The methodology is rigorously applied to:

• Fokker-Planck Dynamics: Continuous space with diffusion $D(x)$ and force $F(x)$.
• Markov Jump Processes: Discrete lattice with finite-difference operators.
  The paper demonstrates that the algebraic structure (tridiagonal matrices in discrete vs. differential operators in continuous) remains consistent under the unified $\nabla$ notation.

4. Variable Transformations for Simplification

The author identifies two specific coordinate systems that simplify the analysis:

• Variable $y$: Defined such that $y - y_1 = L_1(x)$. In this frame, the slowest observable is linear ($L_1(y) = y - y_1$), and the Ito force becomes linear. This simplifies the analysis of Pearson-type models (quadratic diffusion).
• Variable $z$: Defined such that the diffusion coefficient is constant ($d(z)=1$). This maps the problem to standard Schrödinger equations with a scalar potential, recovering known results in SUSY QM literature.

4. Key Results

• Explicit Reconstruction Formula: The steady state $P^*(x)$ for a given $L_1(x)$ and $E_1$ is explicitly given by:
  $$P^*(x) = \frac{1}{D(x) L_1'(x) Z} \exp\left( -E_1 \int^x \frac{L_1(X)}{D(X) L_1'(X)} dX \right)$$
  This allows the construction of an infinite family of QES models by simply choosing different functions for $L_1(x)$.
• Relation to Pearson Models: When the diffusion coefficient is a quadratic polynomial and the force is linear (Pearson family), the method naturally recovers the known exactly solvable cases (e.g., Hermite, Laguerre, Jacobi polynomials) as specific instances where $L_1(x)$ is a polynomial.
• Doob Transformation Equivalence: The paper proves that the Doob transformation of the Markov generator $\tilde{G}$ using $\tilde{L}_1(x)$ is mathematically identical to the supersymmetric recursion $H^{[1]} = \tilde{H} - E_1$.
• Discrete Case Analysis: In the Appendix, the author derives the specific tridiagonal matrix forms for the generators and shows how the potentials $U(x)$ and $U_I(x)$ are reconstructed from the discrete left eigenvector $\tilde{L}_1(x)$, providing a complete discrete counterpart to the continuous theory.

5. Significance

• Conceptual Clarity: The paper shifts the paradigm from "solving the Schrödinger equation" to "constructing the dynamics from the observable." This provides a more physical intuition for QES models, linking them directly to relaxation rates and steady-state currents.
• Unification: It unifies the treatment of continuous diffusion and discrete jump processes under a single algebraic framework, highlighting that the underlying structure of QES models is independent of the specific nature of the space (continuous vs. discrete).
• Practical Utility: The proposed method offers a constructive algorithm for generating new QES models. By choosing a specific form for the slowest observable $L_1(x)$, researchers can immediately generate the corresponding potential and steady state, bypassing complex algebraic constraints usually required in SUSY QM.
• Extension to Non-Equilibrium: The conclusion notes that these insights extend to non-equilibrium processes (breaking detailed balance) where a non-vanishing steady current exists, suggesting the utility of the "slowest observable" approach beyond equilibrium statistical mechanics.

In summary, Cécile Monthus demonstrates that the slowest observable is the natural "seed" for constructing quasi-exactly-solvable Markov processes. This perspective not only simplifies the mathematical construction of $N=2$ models but also deepens the physical understanding of the relationship between quantum spectral properties and stochastic relaxation dynamics.