Constrained Langevin approximation for the Togashi-Kaneko model of autocatalytic reactions

TL;DR

This paper analyzes a constrained Langevin approximation of the Togashi-Kaneko model, demonstrating its mathematical properties and validating its effectiveness in approximating the original stochastic reaction network's behavior.

Contribution

The paper establishes the mathematical properties of the CLA for the TK model and shows it accurately approximates the original model's stationary distribution and transition dynamics.

Findings

CLA is a Feller process and positive Harris recurrent

Exponential convergence to stationary distribution

CLA effectively approximates TK model's transition times

Abstract

The Togashi Kaneko model (TK model), introduced by Togashi and Kaneko in 2001, is a simple stochastic reaction network that displays discreteness-induced transitions between meta-stable patterns. Here we study a constrained Langevin approximation (CLA) of this model. The CLA, obtained by Anderson et al. in 2019, is an obliquely reflected diffusion process on the positive orthant and hence it respects the constrain that chemical concentrations are never negative. We show that the CLA is a Feller process, is positive Harris recurrent, and converges exponentially fast to the unique stationary distribution. We also characterize the stationary distribution and show that it has finite moments. In addition, we simulate both the TK model and its CLA in various dimensions. For example, we describe how the TK model switches between meta-stable patterns in dimension 6. Our simulations suggest that, under the classical scaling, the CLA is a good approximation to the TK model in terms of both the stationary distribution and the transition times between patterns.