Beyond Classical Diffusion: Fractional Derivatives in Transport and Stochastic Systems
Cypres Verbeeck, Nikolaos Sfakianakis · 2025
Read Beyond Classical Diffusion: Fractional Derivatives in Transport and Stochastic Systems by Cypres Verbeeck, Nikolaos Sfakianakis (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-2503-13096.txt.
Introduction
Verbeeck and Sfakianakis explored fractional calculus for transport and random systems in twenty twenty-five. Standard calculus handles local, uniform effects well. Complex systems such as living tissue vary across space and remember their past, calling for richer tools.
Fractional derivatives, which generalize the idea of a derivative to non-integer orders, supply those tools. The authors derive fractional reaction-diffusion equations from continuous-time random walks, where heavy-tailed waiting patterns drive the result. They build equations at both the whole-system and single-particle levels and simulate simple cases. A new numerical scheme keeps total mass from leaking under repeating boundary conditions. The paper contrasts fractional spread with classical diffusion, which moves in the familiar bell-curve fashion.
Catalog record
| CREATOR | Cypres Verbeeck, Nikolaos Sfakianakis |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 107 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-13096
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Beyond Classical Diffusion: Fractional Derivatives in Transport and Stochastic Systems Thanks: Submitted to the editors August 9, 2026. Cypres Verbeeck, Nikolaos Sfakianakis Abstract Integer-order differential operators were originally used to describe local and isotropic effects, in both space and time. However, in fields like biology, the modelling of complex phenomena with spatial heterogeneity necessitates more advanced approaches. The fractional calculus framework provides powerful tools for developing models that better capture the intricate dynamics of biological systems. This paper derives fractional reaction-diffusion equations from continuous-time random walks, highlighting the role of heavy-tailed distributions in the process. Both fractional partial differential equations, on the macroscopic level, as well as fractional stochastic differential equations, on the microscopic level, will be derived and simulated from, for simple Riesz-fractional diffusion models. A new numerical scheme that implements periodic boundary conditions is proposed to control the loss of mass density. We highlight the key differences between fractional and classical diffusion. keywords Fractional diffusion; Riesz- and Caputo fractional derivatives; continuous-time random walk; fractional stochastic differential equation. † † runningheads: Beyond Classical Derivatives / C. Verbeeck, N. Sfakianakis AMS 26A33; 35R11; 44Axx; 60G22 1 Introduction Differential operators of integer order are known to capture local and isotropic effects, both in space and time. However, especially in biology, the increasing need to model complex phenomena with underlying properties such as spatial heterogeneity and/or effects of memory requires new (and more complex) modelling tools, Vieira et al., 2023 ; West, 2016 . The fractional calculus framework enables the development of more sophisticated models that accurately capture the complex dynamics inherent in various systems: including viscoelasticity, Mainardi, 2010 , hydrological processes, Benson et al., 2000 , and biological applications such as bioengineering, Magin, 2006 . Fractional calculus could even be applied to describe motor control systems, notably the dynamics of motor and premotor neurons, Anastasio, 1994 . Specifically, we want to see for ourselves how fractional reaction-diffusion equations arise in a natural way, starting from basic principles. Similar to how one would derive the macroscopic diffusion equation, a continuous-time random walk (CTRW), together with mass conservation equations, provide the starting point of our derivations. The use of CTRWs in fractional diffusion modelling has been extensively studied, particularly in the context of numerical simulations of individual particle trajectories, by Kleinhans and Friedrich, 2007 . Scalas et al., 2004 approached fractional diffusion in 2004, from a probabilistic perspective. Inspired by their methodology and results, this paper transfers their ideas to the slightly different density setting, where creation-destruction phenomena are also factored into the system’s progression. Starting from the mesoscopic density equations provided by Méndez et al., 2010 , we are able to argue how fractional derivatives naturally make their appearance in the resulting macroscopic description of the system. Diffusion describes the physical process of multiple particles undergoing a random walk on a microscopic level. Integer-order diffusion equations present the most well-known elementary models to describe the diffusion process. They assume that the stochastic walks follow a standard Gaussian distribution. However, this simplistic underlying assumption neglects to incorporate lesser effects of complicated biological scenarios, which are nonetheless equally important as more prominent (collective) movements. Moreover, there are many diffusion processes in nature that do not satisfy the Gaussianity assumption. Hence, employing the integer-order diffusion equation does not always lead to an accurate representation of diffusion as observed in reality. In porous media, for instance, anomalous diffusion is perceived to be dominant. To quote Xie and Fang, 2022 : “A large number of experiments show that the fractional diffusion equations can provide a more accurate mathematical description of these anomalous diffusion processes in nature than the integer-order diffusion model […]”. For a detailed review of anomalous diffusion and fractional dynamics, see also Metzler and Klafter, 2000 or Henry and Wearne, 2000 . For a broader perspective on fractional dissipative PDEs—including reaction-diffusion systems, their stability, bifurcation and travelling waves analysis, as well as some numerical simulations—we refer to Achleitner et al., 2024 ; Ortigueira, 2013 . Fractional diffusion equations are starting to attract more and more interest. Notably, Riesz-fractional diffusion equations have been adopted in many scientific and engineering applications. Often, an exact analytical solution of one such model is either not available or too complicated to reproduce. This is why numerical methods for the simulation of fractional (diffusion) models need to be developed. We refer the interested reader to Li and Zeng’s Numerical Methods for Fractional Calculus for an overview of numerical methods for the computation of various fractional differential equations, Li and Zeng, 2015 . Recent advances have introduced memory-efficient spectral methods for solving time-fractional PDEs, significantly reducing computational costs while maintaining accuracy, see Gutleb and Carrillo, 2023 . Although various numerical methods have already been introduced for Riesz (one-dimensional) fractional diffusion in particular, Xie and Fang, 2022 , many schemes fail to correctly account for mass conversation. Most available numerical methods assume Dirichlet boundary conditions. See also Nochetto et al., 2015 for a study of solution techniques for problems involving fractional powers of symmetric coercive elliptic operators in a bounded domain with Dirichlet boundary conditions. In Huang and Oberman, 2014 , far-field boundary conditions are assumed to solve the fractional Laplacian numerically. Hermite spectral methods lie at the heart of solving fractional PDEs in unbounded domains in Mao and Shen, 2017 , while a spectral-Galerkin method solves “integral fractional Laplacian PDEs” in unbounded domains of \mathbb{R}^{d} in Sheng et al., 2020 . Nevertheless, plenty of real-life problems involving space-fractional diffusion on a bounded domain require mass-conserving boundary conditions instead, Kelly et al., 2019 . This paper aims to compare fractional against regular diffusion on both the microscopic as well as the macroscopic level. The microscopic model calls for stochastic differential equations, drawing from normal and stable distributions. On the macroscopic level, we propose a new numerical scheme that incorporates periodic boundary conditions, by closing up a one-dimensional spatial domain. By doing so, we avoid any blow-up phenomena that may occur at the boundary of the domain, see also Abatangelo et al., 2023 for an elaborate discussion of the explosive behaviour characteristic of equations driven by non-local—fractional—operators. Existing approximations will form the basis of a modified Grünwald-Letnikov discretisation for the spatial Riesz derivative. Another possible discretisation of the spectral fractional Laplacian on a bounded domain is discussed by, for instance, Cusimano et al., 2018 . Their formalism is naturally made to handle different types of boundary constraints. We use the Grünwald-Letnikov operator approximation at the midpoint of the domain and replicate the resulting stencil at every other lattice point. We will do so by introducing a fixed vector carrying the weights of the sum, all the while permuting the position indices of all spatial grid points, to account for ghost points and enforce symmetry of the scheme itself. The periodicity acts to transfer information across boundaries and was introduced to control the mass in finite domains. Whilst succeeding in controlling the loss of mass at the endpoints, a complete conservation of mass is not observed, for neither regular nor fractional (of order 1.5 ) diffusion. Of course, the choice of different parameters, such as final time T , temporal step size \tau , and size of the domain itself, affect the results quantitatively. Larger