Traveling waves in Bistable Reaction-Diffusion Cellular Automata

Daniel Špale, Petr Stehlík · 2024

Read Traveling waves in Bistable Reaction-Diffusion Cellular Automata by Daniel Špale, Petr Stehlík (2024) 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-2412-17441.txt.

Introduction

Waves that travel through a material with two stable states appear in chemistry, biology, and physics. Daniel Špale and Petr Stehlík study fully discrete versions of such systems in a 2024 paper, where time, space, and state all jump in steps, as in a grid of simple cells. They sort the kinds of moving fronts these systems can produce.

With strong spreading, waves move only slowly and their shapes are not unique. With weak spreading, waves can freeze in place. Between those extremes sit unusual waves, found only in the discrete setting, that move while their shape repeats in a cycle. The authors map where each kind occurs.

Catalog record

CREATORDaniel Špale, Petr Stehlík
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE49 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-17441

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Traveling waves in Bistable Reaction-Diffusion Cellular Automata Daniel Špale, Petr Stehlík Abstract . We describe various types of traveling fronts of bistable reaction-diffusion cellular automata. These dynamical systems with discrete time, space, and state spaces can be seen as fully discrete versions of widely studied bistable reaction-diffusion equations. We show that moving traveling waves for high diffusion parameters are restricted to slow speeds and their profiles are interestingly not unique. Pinned waves always exist for weak diffusion as in the case of lattice equations but do not complement parametric region of moving traveling waves. The remaining parameter domain is dominated by waves which are unique to cellular automaton settings. These higher-order traveling waves move and periodically change profile at the same time. 2020 Mathematics Subject Classification Primary 37B15; Secondary 34A33, 39A12 1. Introduction Waves in lattice reaction-diffusion equations provide one of the most fascinating distinctive behavior between the dynamics of spatiotemporal dynamical systems which differ only in underlying discrete and continuous structures. As the most typical example, the Nagumo partial differential equation (i.e., continuous space) serves as a textbook example for the phase plane analysis which provides existence of traveling waves [ 15 ] . In contrast, the Nagumo lattice differential equation (discrete space) is far more complicated to study since it yields several new dynamical phenomena, e.g., pinning of waves [ 14 ] , spatial topological chaos [ 3 ] , and existence of nonmonotone waves [ 13 ] . In this paper, we study traveling fronts of a class of bistable reaction-diffusion cellular automata, which correspond to Nagumo partial and lattice differential equations but are fully discrete, i.e., temporal, spatial, and state sets are all discrete. In Section 2 , we introduce a bistable discrete dynamical system (1.1) u(t+1)=F(u(t)),\quad t\in{\mathbb{N}}_{0},\ u(t)\in S^{\mathbb{Z}}, with the finite state space S\subset{\mathbb{N}}_{0} and the mapping F:S^{\mathbb{Z}}\rightarrow S^{\mathbb{Z}} which involves both diffusion and reaction processes. We then compare its behavior to the widely studied corresponding partial and lattice differential equations and study three types of fronts - moving, pinned, and periodically changing higher-order traveling waves. Beyond the natural interest in distinct mathematical properties, our motivation is driven by several applications where discrete time, space, and states are common, e.g., seasonal populations (discrete time, [ 2 ] ), island or pond populations (discrete space, [ 12 ] ), or very small populations (discrete states, [ 8 ] ). (Nonspatial) bistable dynamics. Various biological, chemical and mechanical systems possess two stable equilibria. The simplest models are connected to ordinary differential models with cubic nonlinearities, e.g., u^{\prime}(t)=u(t)\left(1-u^{2}(t)\right),\quad t\geq 0, with stable equilibria u^{*}_{1,2}=\pm 1 and the unstable equilibrium u_{3}^{*}=0 . In mathematical biology, it is more natural to generalize the standard logistic growth with a capacity K>0 into the bistable models with the so called Allee effect. Roughly speaking, the (demographic) Allee effect captures phenomena in which populations smaller than a viability threshold a with 0<a<K go extinct. Mathematically, this leads to the bistable model (1.2) u^{\prime}(t)=\lambda u(t)\left(\frac{u(t)}{A}-1\right)\left(1-\frac{u(t)}{K}\right),\quad t\geq 0, with the stable equilibria u^{*}_{1}=0 (extinction), u^{*}_{2}=K (carrying capacity) and unstable equilibrium u_{3}^{*}=A (viability threshold). Numerous examples from applications, list of alternative mathematical approaches, and deep discussions about the Allee effect and nonspatial bistable dynamics can be found in [ 7 ] . Waves in bistable partial differential equations. Nondimensionalisation of ( 1.2 ) and the use of the standard continuous spatial diffusion leads to the widely-studied Nagumo bistable partial differential equation (PDE) (1.3) u_{t}(x,t)=du_{xx}(x,t)+u(x,t)\left(u(x,t)-a\right)\left(1-u(x,t)\right),\quad t\geq 0,\quad x\in\mathbb{R}, where d>0 denotes the diffusion parameter. Initially, this equation has been considered as a simplified model for the pulse propagation in nerve axons [ 19 ] . The PDE has a traveling wave solution which connects the two stable equilibria u^{*}_{1}=0 and u^{*}_{2}=1 [ 15 ] (1.4) u(x,t)=\frac{1}{2}+\frac{1}{2}\tanh\left(\frac{x-ct}{2\sqrt{2d}}\right), where c denotes the speed of propagation and satisfies c=\sqrt{2d}\left(a-\frac{1}{2}\right), i.e., the wave profile ( 1.4 ) propagates left whenever a<1/2 , right for a>1/2 , and is pinned when a=1/2 , see the left panel of Fig. 1 . Figure 1 . Comparison of properties of traveling waves in (a,d) -plane for bistable partial differential equation (left panel), lattice differential equation (central panel) and cellular automaton (right panel). Dark regions represent left or right moving traveling waves, light regions pinned waves and middle gray regions correspond to higher-order waves which are unique to cellular automata. Waves in bistable lattice differential equations. In many applications (sequence of cells, neurons, spatial habitats) the continuous space X=\mathbb{R} in ( 1.3 ) is naturally replaced by an integer lattice X=\mathbb{Z} and we obtain the Nagumo lattice equation (1.5) u^{\prime}_{i}(t)=d\left(u_{i-1}(t)-2u_{i}(t)+u_{i+1}(t)\right)+u_{i}(t)\left(u_{i}(t)-a\right)\left(1-u_{i}(t)\right),\quad t\geq 0,\quad i\in\mathbb{Z}. The same equation can be obtained from the PDE ( 1.3 ) by applying standard central difference scheme to the second spatial derivative. The dynamical properties of the LDE ( 1.5 ) differ significantly from those of the PDE ( 1.3 ). There is no explicit solution like ( 1.4 ). For any a\in(0,1) and sufficiently small diffusion d the waves are pinned with c=0 and there exists large number of stationary solutions [ 14 ] . For any d>0 there exists \varepsilon>0 such that c=0 whenever \left\lvert a-1/2\right\rvert<\varepsilon , see [ 3 ] and the central panel in Fig. 1 . The exact boundaries between c=0 and c\neq 0 remain unknown and the details of the pinning procedures have been intensively studied in simplified bistable caricatures [ 10 ] . Alternative propagation schemes have been observed in the pinning regions, e.g., nonmonotone multichromatic waves [ 13 ] . Similar pinning behavior has been studied as well in the models when also time is considered to be discrete (or discretized by the forward Euler discretization scheme), e.g., [ 5 ] , (1.6) u_{i}(t+1)-u_{i}(t)=d\left(u_{i-1}(t)-2u_{i}(t)+u_{i+1}(t)\right)+u_{i}(t)\left(u_{i}(t)-a\right)\left(1-u_{i}(t)\right),\quad t\in\mathbb{N}_{0},i\in\mathbb{Z}. Patterns and waves in bistable cellular automata. There have been numerous attempts to study patterns and waves in bistable cellular automata, i.e., models in which not only time and space are discrete but also the values of u are reduced to a finite number of states. In many cases, the focus has been on the discretization techniques, standard finite difference schemes [ 24 ] , the tropical discretization [ 18 ] , or the fine correspondence of three-letter state set model with the behavior of lattice equations ( 1.5 )–( 1.6 ), [ 4 ] . The natural presence of propagation phenomena has been identified in reaction-diffusion cellular automata by techniques ranging from simulations [ 9 ] to a systematic combinatorial analysis [ 16 ] . Numerous wave types include periodic waves [ 1 , 6 , 21 ] and spiral waves [ 9 ] based on the famous excitable cellular automata by Greenberg and Hastings [ 11 ] . Large number [ 23 ] and numerous types [ 4 ] of stationary patterns of reaction-diffusion automata have also been observed in reaction-diffusion automata. Higher-order waves. In this paper we focus on a class of bistable cellular automata and compare their propagation properties to those of the PDE ( 1.3 ) and the LDE ( 1.5 ). We show that beyond monotonic moving traveling waves and pinned waves there exists an intermediate region in which there are no moving or pinned traveling waves. The dynamics in this region are driven by a new type of waves which

Read the complete text