From random walks on networks to nonlinear diffusion

TL;DR

This paper connects random walks on networks to nonlinear diffusion equations, revealing how network structure influences diffusion behavior and providing insights into complex biological and physical systems.

Contribution

It demonstrates that in the continuum limit, stochastic random walks on networks correspond to nonlinear diffusion PDEs with density-dependent coefficients, extending understanding of diffusion processes.

Findings

Diffusion coefficient relates to transition probabilities.

Relaxation time depends on network structure.

Self-similar solutions show finite propagation speed.

Abstract

Mathematical models of motility are often based on random-walk descriptions of discrete individuals that can move according to certain rules. It is usually the case that large masses concentrated in small regions of space have a great impact on the collective movement of the group. For this reason, many models in mathematical biology have incorporated crowding effects and managed to understand their implications. Here, we build on a previously developed framework for random walks on networks to show that in the continuum limit, the underlying stochastic process can be identified with a diffusion partial differential equation. The diffusion coefficient of the emerging equation is in general density-dependent, and can be directly related to the transition probabilities of the random walk. Moreover, the relaxation time of the stochastic process is directly linked to the diffusion coefficient and also to the network structure, as it usually happens in the case of linear diffusion. As a specific example, we study the equivalent of a porous-medium type equation on networks, which shows similar properties to its continuum equivalent. For this equation, self-similar solutions on a lattice and on homogeneous trees can be found, showing finite speed of propagation in contrast to commonly used linear diffusion equations. These findings also provide insights into reaction-diffusion systems with general diffusion operators, which have appeared recently in some applications.