Multi-parameter isospectral Fokker-Planck equations
Choon-Lin Ho · 2025
Read Multi-parameter isospectral Fokker-Planck equations by Choon-Lin Ho (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-12939.txt.
Introduction
Choon-Lin Ho builds whole families of Fokker-Planck equations that share the same spectrum, in a 2025 mathematical physics paper. Starting from one such equation, which governs how probability spreads under drift and noise, Ho deletes a chosen set of its eigenstates with a multi-step Darboux-Crum transformation. The deleted energy levels are then put back by running the transformation in reverse, landing on a deformed partner equation with the same levels.
A brief extension to fractional versions of the equation is included. The paper also comments on a recent one-parameter version of the trick applied to black holes in a thermal potential picture. The method gives many distinct-looking systems with identical spectra.
Catalog record
| CREATOR | Choon-Lin Ho |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 28 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-12939
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Multi-parameter isospectral Fokker-Planck equations Choon-Lin Ho Abstract From a given Fokker-Planck equation, a multi-parameter deformed partner Fokker-Planck equation is constructed. This is done by first deleting a set of eigenstates of the original FPE by the multi-step Darboux-Crum transformation, and then reinstating the eigen-energy levels by the reverse Darboux-Crum transformation. Extension to fractional Fokker-Planck equation is briefly discussed. A recent study of the one-parameter isospectral FPE applied to black hole in the thermal potential approach is commented. pacs 05.10.Gg; 02.30.Ik; 05.90.+m; 02.50.Ey I Introduction The Fokker-Planck equation (FPE) is one of the basic equations used to study various phenomena that involve fluctuations and dissipation Ris ; Sau . As such phenomena are quite common in nature, it is therefore not surprising that FPE has found applications in such diverse areas as physics, astrophysics, chemistry, biology, finance, etc. As with any equation we encounter, it is generally not easy to find analytic solutions of the FPE. Exact analytical solutions of the FPEs are known for only a few cases, such as linear drift and constant diffusion coefficients Ris ; Sau . One way to solve the FPE with time-independent drift and diffusion coefficients is to transform the FPE into a stationary Schrödinger equation, and then solve the eigenvalue problem of the latter. Several FPEs have been exactly solved in this way when the associated schrödinger equations are solvable Ris ; Sau . Recently, Such connection has been employed to investigate quasi-exactly solvable FPEs ho1 , and FPEs with solutions based on the newly discovered exceptional orthogonal polynomials ho2 ; ho3 . Furthermore, such correspondence between FPE and Schrödinger equation allows one to exploit the formalism of supersymmetric (SUSY) quantum mechanics SUSY ; Sa . In SUSY quantum mechanics, the spectrum of a Schrödinger equation can be related to that of another through the Darboux transformation Dar ; Cru . The SUSY method was first employed in BB to compute the small eigenvalue of the Fokker-Planck equation which controls the rate at which equilibrium is approached. Since then there appear many studies employing SUSY method in various ways to study the FPE, see eg., K ; SMJ ; Sch ; IN ; ho4 ; ho5 and references therein. Particularly, solutions of FPEs related by a one-parameter isospectral SUSY transformation Le were studied in Ros ; R ; P . In Ref. ho4 we have made use of the Darboux transformation to find solutions of a FPE from the solutions of another FPE. FPEs with either time-independent or dependent drifts (and constant diffusion) were considered. In the case of time-independent drift, we construct a hierarchy of Darboux transform-related FPEs using the property of shape invariance. Shape invariance is a sufficient condition for exact solvability, which amazingly is possessed by most of the exactly-solvable one-dimensional quantum systems. In ho5 we consider a family of one-parameter deformation from one FPE to its SUSY partner FPE. In this work we would like to consider relating two FPEs with time-independent drift and constant diffusion by multi-parameter isospectral transformation Le , generalizing the work in Ros , which considers a one-parameter transformation. By multi-parameter isospectral transformation we mean that the transformed FPE has the same eigenvalues as the original one but with different forms of eigenstates. This is done by first deleting a set of eigenstates of the original FPE by the multi-step Darboux tranformation, which is usually called the Darboux-Crum transform, and then reinstating the eigen-energy levels by the reverse Darboux transformation. We will briefly review the connection between FPE and Schrödinger equation in Sect. II. Sect. III explains the main ideas of construction of FPE pair based on the Darboux-Crum transformation. Multi-parameter Isopectral FPEs are discussed in Sect. IV. These procedures are then extended to a class of fractional FPE in Sect. V. In the last section we summarize the paper, and briefly comment on a recent attempt to apply the isopectral transformation to black holes in thermal potential. II FPE and Schrödinger equation The FPE with constant diffusion coefficient (which we set to be unity without loss of generality) and time-independent drift D^{(0)}(x) is \frac{\partial}{\partial t}P(x,t)=\left[-\frac{\partial}{\partial x}D^{(0)}(x)+\frac{\partial^{2}}{\partial x^{2}}\right]P(x,t). (1) The domain of x is taken to be x\in[c_{1},c_{2}] . Typical values of c_{1},c_{2} are c_{1}=0,-\infty and c_{2}>0,+\infty . It is understood that, as probability density function (PDF), P(x,t) is to be normalized over the domain of x . It is convenient to write D^{(0)}(x)=-2W^{\prime}(x)\equiv-U^{\prime}(x) , where the prime denotes x -derivative. The function U(x)\equiv 2W(x) is called the drift potential of the system. Putting P(x,t)=e^{-W(x)}\exp(-\epsilon t)\phi(x) (2) in ( 1 ), we obtain \displaystyle~~~~~H\phi=\epsilon\phi, (3) \displaystyle H \displaystyle\equiv \displaystyle-\partial_{x}^{2}+W^{\prime 2}-W^{\prime\prime}. For clarity of presentation, we often left out arguments in formulas in what follows. The function \phi satisfies the stationary Schrödinger equation with Hamiltonian H and eigenvalue \epsilon . The Hamiltonian H can be written as H=A^{(0)\dagger}A^{(0)} , with the pair of Hermitian conjugate operators A^{(0)}=\partial_{x}+W^{\prime},~~A^{(0)\dagger}=-\partial_{x}+W^{\prime}. (4) Taking the overlap integral of \phi with the first equation in ( 3 ), we have \epsilon\langle\phi|\phi\rangle=\langle\phi|H\phi\rangle=\langle\phi|A^{(0)\dagger}A^{(0)}\phi\rangle=\langle A^{(0)}\phi|A^{(0)}\phi\rangle\geq 0. (5) Since \langle\phi|\phi\rangle>0 for non-vanishing states, the above relation then implies that the eigenvalues are non-negative, i.e. \epsilon\geq 0 . We will arrange \epsilon in increasing order : 0\leq\epsilon_{0}<\epsilon_{1}<\epsilon_{2}<\ldots . The function \phi_{0}\equiv\exp(-W) is the zero mode of H : H\phi_{0}=0 . If \phi_{0} is non-normalizable, then it is not an eigenstate of H , and \epsilon_{0}=0 is absent in the spectrum. For definiteness, we consider only the case when \phi_{0} is an eigenstate. In this case, W=-\ln|\phi_{0}| , and A^{(0)}=\partial_{x}-(\ln|\phi_{0}|)^{\prime},~~A^{(0)\dagger}=-\partial_{x}-(\ln|\phi_{0}|)^{\prime}. (6) Note that A^{(0)}\phi_{0}=0 . The drift is related to the ground state by D^{(0)}(x)=2(\ln|\phi_{0}|)^{\prime} . We mention here that the construction to be discussed below can be easily extended to the case where \epsilon_{0}=0 is absent with slight modification using \phi_{1} in place of \phi_{0} . This connection of FPE with the Schrödinger equation allows the FPE be solved by the method of eigenfunction expansion. If all the eigenfunctions \phi_{k} ( k=0,1,2,\ldots ) of H with eigenvalues \epsilon_{k} are solved, then the solution of P_{k}(x,t) of ( 1 ) corresponding to the eigenvalue \epsilon_{k} is P_{k}(x,t)=\phi_{0}(x)\phi_{k}(x)\exp(-\epsilon_{k}t) . The stationary distribution is P_{0}(x)=\phi_{0}^{2}=\exp(-2W) , which is obviously non-negative. Any positive definite initial probability density P(x,0) can be expanded as P(x,0)=\phi_{0}(x)\sum_{k}c_{k}\phi_{k}(x) , with constant coefficients c_{k} ( k=0,1,\ldots ) \displaystyle c_{k}=\int_{-\infty}^{\infty}\phi_{k}(x)\left(\phi_{0}^{-1}(x)P(x,0)\right)dx. (7) Then at any later time t , the solution of the FP equation is P(x,t)=\phi_{0}(x)\sum_{k}c_{k}\phi_{k}(x)\exp(-\epsilon_{k}t). (8) In this work we consider how to obtain solutions of a FPE from those of another by the Darboux-Crum and isospectral transformations between the two Schrödinger equations associated with them. III Darboux-Crum transformation In Sch the construction of Darboux-Crum pair of FPEs is considered by deleting a set of auxiliary states of the Schrödinger equation. Since our purpose here is the isospectral FPEs, we shall consider only Darboux-Crum transformation using the lowest lying eigenstates. To make our presentation self-contained, and to fix notations, we shall briefly review this transformation based on the set of lowest lying states Cru ; Sa . As explained in the last section, the Hamiltonian H can be written as H=A^{(0)\dagger}A^{(0)} . The Darboux theorem states that the Hamiltonian H^{(1)}=A^{(0)}A^{(0)\dagger} has (unnormalized) eigenstates \phi_{k}^{(1)}=A^{(0)}\phi_{k},~~A^{(0)}\phi_{0}=0,~~~k=1,2,\ldots, (9) with eigenvalues \{\epsilon_{1},\epsilon_{2},\ldots\} . Thus H^{(1)} has the same spectrum as H