Heterogeneous diffusion process with power-law nonlinearity
Jorge E. Cardona, Ilya Pavlyukevich · 2025
Read Heterogeneous diffusion process with power-law nonlinearity by Jorge E. Cardona, Ilya Pavlyukevich (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-2512-01828.txt.
Introduction
Jorge E. Cardona and Ilya Pavlyukevich study a random motion whose speed depends on where the moving particle currently sits. Their 2025 paper treats a diffusion process with a power-law nonlinearity, written as a stochastic differential equation driven by Brownian motion, the standard model of random wandering.
Because the noise term grows or shrinks with position, the particle spreads in ways that plain Brownian motion does not, and standard tools can break down. The authors work out how solutions of the equation behave under this position-dependent randomness. Results of this kind matter wherever spread is uneven, from particles in varied media to models in finance.
Catalog record
| CREATOR | Jorge E. Cardona, Ilya Pavlyukevich |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 74 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-01828
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Heterogeneous diffusion process with power-law nonlinearity Jorge E. Cardona, Ilya Pavlyukevich Abstract In this paper, we study solutions of the heterogeneous diffusion process with power-law nonlinearity governed by the stochastic differential equation d X t = | X t | α d B t + α λ | X t | 2 α − 1 sign ( X t ) d t \mathrm{d}X_{t}=|X_{t}|^{\alpha}\,\mathrm{d}B_{t}+\alpha\lambda|X_{t}|^{2\alpha-1}\sign(X_{t})\,\mathrm{d}t , where α ∈ ( 0 , 1 ) \alpha\in(0,1) and λ ∈ [ 0 , 1 ] \lambda\in[0,1] . The parameter α \alpha controls the nonlinear power-law profile of the diffusion coefficient, while the parameter λ \lambda specifies the interpretation of the stochastic integral in the pre-equation X ˙ = | X | α B ˙ \dot{X}=|X|^{\alpha}\dot{B} . We demonstrate that the solutions of this equation can be represented as nonlinear transformations of a skew Bessel process with dimension δ ∈ ℝ \delta\in\mathbb{R} . Keywords: Bessel process; skew Bessel process; Itô stochastic integral; Stratonovich stochastic integral; Hänggi–Klimontovich interpretation; random time change. 2010 Mathematics Subject Classification: 65C30 ∗ Stochastic differential and integral equations; 60H10 Stochastic ordinary differential equations; 1 Introduction In the physical literature, a Brownian motion process B is often referred to as free diffusion , while the solution of the SDE \displaystyle\dot{X} \displaystyle=\sigma(X)\dot{B} (1.1) is called a heterogeneous diffusion , with \sigma representing the diffusion coefficient. From a mathematical perspective, this corresponds to a one-dimensional SDE with multiplicative noise. However, from the physical point of view, equation ( 1.1 ) does not completely specify the process X . Instead, it serves as a “pre-equation” that must be complemented by an interpretation rule in order to define an actual stochastic process (see van Kampen 1981 ). The interpretation rule determines the choice of stochastic integral. In the classical Itô framework, the SDE ( 1.1 ) is interpreted as the stochastic integral equation \displaystyle X_{t}=x+\int_{0}^{t}\sigma(X_{s})\,\mathrm{d}B_{s} \displaystyle:=x+\lim_{n\to\infty}\sum_{k}\sigma(X_{t_{k}^{n}\wedge t})(B_{t_{k+1}^{n}\wedge t}-B_{t_{k}^{n}\wedge t}) (1.2) where \{t^{n}_{k},k\in\mathbb{N}_{0}\}_{n\in\mathbb{N}} denotes a sequence of partitions of the half-line [0,\infty) with mesh size tending uniformly to zero, i.e., \lim_{n\to\infty}\max_{k}|t^{n}_{k+1}-t^{n}_{k}|=0 , The limit in ( 1.2 ) is taken in the ucp (uniform convergence on compacts in probability) sense, see, for example, Chapter II in Protter 2004 . It is the characteristic feature of the Itô stochastic integral, that the integrand \sigma(X) is always evaluated at the left endpoint of each partition interval. Different interpretations of multiplicative noise can be introduced by means of an interpretation parameter \lambda\in[0,1] . The corresponding \lambda -stochastic integral is defined as the ucp-limit \displaystyle\int_{0}^{t}\sigma(X_{s})\circ_{\lambda}\mathrm{d}B_{s} \displaystyle:=\lim_{n\to\infty}\sum_{k}(\lambda\sigma(X_{t_{k+1}^{n}\wedge t})-(1-\lambda)\sigma(X_{t_{k}^{n}\wedge t}))(B_{t_{k+1}^{n}\wedge t}-B_{t_{k}^{n}\wedge t}), (1.3) which can be rewritten as \displaystyle\int_{0}^{t}\sigma(X_{s})\circ_{\lambda}\mathrm{d}B_{s} \displaystyle=\int_{0}^{t}\sigma(X_{s})\,\mathrm{d}B_{s}+\lambda[\sigma(X),B]_{t}, (1.4) where [\sigma(X),B] denotes the bracket process \displaystyle{}[\sigma(X_{t}),B]_{t}:=\lim_{n\to\infty}\sum_{k}(\sigma(X_{t_{k+1}^{n}\wedge t})-\sigma(X_{t_{k}^{n}\wedge t}))(B_{t_{k+1}^{n}\wedge t}-B_{t_{k}^{n}\wedge t}), (1.5) provided, the limit exists. If \sigma is sufficiently smooth, for instance, \sigma\in C^{2} , the bracket process can be expressed explicitly as a Lebesgue integral, \displaystyle{}[\sigma(X),B]_{t}=\int_{0}^{t}\sigma(X_{s})\sigma^{\prime}(X_{s})\,\mathrm{d}s. (1.6) In this case, the interpretation parameter \lambda contributes an additional noise-induced drift term \lambda\sigma(X)\sigma^{\prime}(X) , so that the \lambda -interpretation of ( 1.1 ) coincides with the Itô SDE \displaystyle X_{t}=x+\int_{0}^{t}\sigma(X_{s})\,\mathrm{d}B_{s}+\lambda\int_{0}^{t}\sigma(X_{s})\sigma^{\prime}(X_{s})\,\mathrm{d}s. (1.7) The best-known special case is the Stratonovich interpretation with \lambda=\frac{1}{2} ; see Section V.5 of Protter 2004 . Another important case is \lambda=1 , referred to in the physics literature as the Hänggi–Klimontovich or kinetic interpretation; see Sokolov 2010 . This paper focuses on a particular heterogeneous diffusion with the irregular diffusion coefficient \sigma(x)=|x|^{\alpha} , \alpha\in(0,1) . Such a diffusion was first considered in the Itô setting by Girsanov 1962 as an example of an SDE without the uniqueness property. More recently, it was examined in Cherstvy et al. 2013 in the context of the Stratonovich SDE \displaystyle\dot{X}=|X|^{\alpha}\circ\,\dot{B}, (1.8) corresponding to the interpretation parameter \lambda=\tfrac{1}{2} . In that work, the authors approached the problem at a formal, physical level of rigour and derived the probability density function for a symmetric solution. Further, in Chapter 3.3 of Heidernätsch 2015 and in Kazakevičius and Ruseckas 2016 , explicit densities of non-negative solutions for general \lambda -interpretations were presented. The Stratonovich heterogeneous diffusion ( 1.8 ) was studied rigorously using stochastic calculus in Pavlyukevich and Shevchenko 2020 and Pavlyukevich and Shevchenko 2025 for \alpha\in(-1,1) . It was shown that the equation is underdetermined and admits infinitely many strong solutions. Restricting attention to the class of homogeneous strong Markov solutions that spend zero time at 0 , one obtains solutions as nonlinear transformations of a skew Brownian motion. Specifically, for \alpha\in(0,1) and any \theta\in[-1,1] , the process \displaystyle X_{t}=|(1-\alpha)Z^{\theta}_{t}|^{\frac{1}{1-\alpha}}\sign(Z^{\theta}_{t}) (1.9) is a strong solution to ( 1.8 ) where Z^{\theta} is the unique strong solution to the SDE \displaystyle Z^{\theta}_{t}=z+B_{t}+\theta L^{0}_{t}(Z^{\theta}),\quad z=\frac{1}{1-\alpha}|x|^{1-\alpha}\sign(x), (1.10) where L^{0} denotes a symmetric semimartingale local time at 0 . For \alpha\in(-1,0] , only the symmetric solution with \theta=0 is possible (see Theorem 4.5 in Pavlyukevich and Shevchenko 2020 ). Moreover, Pavlyukevich and Shevchenko 2025 established that the symmetric solution with \theta=0 , first identified in Cherstvy et al. 2013 , is physically relevant in the sense that it arises as the unique solution of a stochastic selection problem. For \alpha\in(-1,0] , only the symmetric solution with \theta=0 is possible (see Theorem 4.5 in Pavlyukevich and Shevchenko 2020 ). Moreover, Pavlyukevich and Shevchenko 2025 established that the symmetric solution with \theta=0 , first identified in Cherstvy et al. 2013 , is physically relevant in the sense that it arises as the unique solution of a stochastic selection problem. \displaystyle X_{t}=x+\int_{0}^{t}|X_{s}|^{\alpha}\,\mathrm{d}B_{s}+\alpha\lambda\int_{0}^{t}|X_{s}|^{2\alpha-1}\operatorname{sign}(X_{s})\,\mathrm{d}s. (1.11) which is the formal analogue of ( 1.7 ) with \sigma(x)=|x|^{\alpha} . It should be emphasized that while the identity ( 1.6 ) holds for smooth \sigma , the corresponding relation \displaystyle{}[|X|^{\alpha},B]_{t}=\alpha\int_{0}^{t}|X|^{2\alpha-1}\operatorname{sign}(X_{s})\,\mathrm{d}s (1.12) remains an open question in this singular setting. Partial progress in this direction has been obtained in Pavlyukevich and Shevchenko 2020 and Pavlyukevich and Shevchenko 2025 . It is clear that for any initial value x\neq 0 , the SDE ( 1.11 ) admits a unique local solution. Since both the diffusion coefficient and the drift grow sublinearly at infinity, this local solution does not blow up and can be uniquely extended up to the first hitting time of the origin, \displaystyle\tau_{0}=\begin{cases}\inf\{t\in[0,\infty)\colon X_{t}=0\},\\ +\infty,\quad X_{t}\neq 0,\quad t\in[0,\infty).\end{cases} (1.13) Our analysis of the SDE ( 1.11 ) is based on a nonlinear transformation of the solution X X and its reduction to a Bessel process. Such transformations have been employed for heterogeneous diffusions in ( Heidernätsch 2015 , Chapter 3.3) and Carr and Linetsky 2006 in the context of financial mathematics. The intuition is as follows. Consider the nonlinear, one-to-one mapping \displaystyle H_{\alpha}(x)=\frac{1}{1-\alpha}|x|^{1-\alpha}\operatorname{sign}(x),\quad x\in\mathbb{R}, (1.14) with the inverse \displaystyle H^{-1}_{\alpha}(z)=((1-\alpha)|z|)^{\frac{1}{1-\alpha}}\operatorname{sign}(z),\quad z\in\mathbb{R}, (1.15) which appeared earlier in ( 1.9 ). Applying Itô’s formula to the local solution X of ( 1.11 ) with initial value x\neq 0 , we find that the process Z:=H_{\alpha}(X) satisfies \displaystyle Z_{t} \displaystyle=H_{\alpha}(x)+\frac{\delta-1}{2}\int_{0}^{t}\frac{\mathrm{d}s}{Z_{s}}+B_{t},\quad t\in[0,\tau_{0}), (1.16) \displaystyle\delta \displaystyle=\delta_{\alpha,\lambda}=\frac{1-2\alpha(1-\lambda)}{1-\alpha}\in\mathbb{R}. For x\in(0,\infty) , the process Z is a well-known \delta -dimensional Bessel process. If the initial point is negative, then the process \widetilde{Z}=-H_{\alpha}(X) is a Bessel process satisfying \displaystyle\widetilde{Z}_{t}=|H_{\alpha}(x)|+\frac{\delta-1}{2}\int_{0}^{t}\frac{\mathrm{d}s}{\widetilde{Z}_{t}}+\widetilde{B}_{t},\quad t\in[0,\tau_{0}), (1.17) with \widetilde{B}=-B . Since we consider solutions on the time interval [0,\tau_{0}) , both X and Z are well-defined and unique up to the first hitting time of the origin. Summarizing this intuition, we see that