Differential method and homotopy analysis method are used for solving the two-dimensional reaction-diffusion model. And the structure of the solutions is analyzed. Finally, the homotopy series solutions are simulated with the mathematical software Matlab, so the Turing patterns will be produced. Overall analysis and experimental simulation of the model show that the different parameters lead to different Turing pattern structures. As time goes on, the structure of Turing patterns changes, and the final solutions tend to stationary state.
Differential Method Homotopy Analysis Method Reaction-Diffusion Model Turing Patterns1. Introduction
In time or space, patterns have nonuniform macroscopic structure with regularity. From the thermodynamic point of view, the nature of the pattern formation can be divided into two categories. One is presenting in the thermodynamic equilibrium conditions, such as the crystal structure of inorganic chemistry, the self-organic pattern formation of organic polymers and so on. The other is the station far from the thermodynamic equilibrium conditions, such as the ripples of the sea, the surface patterns of the animal, the strip clouds in the sky and so on.
Reaction-diffusion system is one of the fundamental equations which describe the motion of the nature. It not only has a wide practical background, but also is used in many fields, for example, predator-prey model, spread of infectious diseases, migration of population and spread of forest fires. Its mathematical model is a special kind of parabolic partial differential equations. As for reaction-diffusion systems, the coupling of nonlinear dynamical and linear diffusion leads to spontaneously producing a variety of ordered or disordered pattern of the system. This is the pattern dynamics of the reaction-diffusion systems [1] . According to the pattern dynamics, the ordered patterns can be divided into two categories: stationary state (such as Turing patterns) and traveling wave (such as spiral pattern). As for the Turing patterns of reaction-diffusion systems, in 1952, Turing [2] for the first time showed that the homogeneous system will be stable with little disturbance for the absence of proliferation, while it becomes unstable with the spatial disturbance for joining the steady-state diffusion. This is the Turing instability of reaction-diffusion equations.
The classical method to study the Turing patterns of reaction-diffusion is the analysis of linear stability method [3] - [8] . Firstly, the nonlinear reaction-diffusion model is turned into linear model. Then, perturbation theory is utilized to study its solution which can be seen as the little perturbation of equilibrium state (i.e. uniform and stable state). And we can obtain the conditions of generating Turing pattern through studying the linear equations with perturbation by the analysis of stability method. Finally, through the numerical simulation, the reaction-diffusion model can obtain different Turing pattern structures. This method shows the relationship between the parameters of generating Turing pattern. But for the strong nonlinear problems, this method may not be applicable.
The range of parameters of the Turing pattern can be obtained by the analysis of linear stability method. Based on the parameters which limit to the range, this article solves the reaction-diffusion model by the use of the combination of differential method and homotopy analysis method. Then the changes of the mechanism of Turing patterns can be under control by the simulation of experimental data and analysis of utilizing this method. Finally, the study of the concrete case shows the feasibility and effectiveness of this method.
2. Homotopy Analysis Solution of Reaction-Diffusion Model
The general mathematical representation of two-dimensional reaction-diffusion model is as follows,
Their definition are as follows: u and v are different reactant concentrations vector; is a control parameter; f and g represent the nonlinear dynamics function of the system; and describe the diffusion coefficient; is Laplace operator; describes the regional space of the research for the model.
Discrete the reaction-diffusion model in discrete nodes . The discretization of the model takes the following form,
And,
According to the thought of homotopy analysis method, we take the value
and impose the initial guess solutions and according to the initial and boundary conditions. Therefore, and are called the zero-order approximate solution of problem (2). Introduce the embedded variable. Auxiliary adjusting parameters meet the conditions and, while auxiliary control functions meet the conditions and. For any, zero-order deformation equation is as follows:
The problem (8) shows that two cases. One is, that; the other is, that. The solutions of zero-order equations and
are continuous from initial guess solutions and to the solutions (i.e. and) of problem (2), while embedded variable q is continuous changing from 0 to 1.
and express the form of Taylor expansion as follows,
where
, ,
The two sides of zero-order Equations (8) are solved m-order derivative about q and divided at the same time. Take can obtain the m-order deformation equation as follows,
where
The solutions of problem (10) are,
Choose the auxiliary function, then
In order to make Taylor series are convergent at, we can choose the value of. Obtain the solutions of problem (2) as follows,
In summary, the m-order homotopy approximate solutions of problem (1) are
3. Solving of Example and Analysis3.1. Homotopy Analysis of Brusselator Reaction Diffusion Model
The typical Brusselator reaction diffusion model is as follows,
where are control parameters for the system; and represent the coefficient of diffusion and; h describes the space step (i.e. the step of x and y are the same length h) and. The uniform steady-state solution to the reaction-diffusion system is. By division, choose the value and. Combined the significance of reaction-diffusion mod-
el and the steady-state solution of the homogeneous perturbation analysis, the guess of initial solution can be as follows,
where is a perturbation parameter and and represent the wave number of x-direction and y-direc- tion, respectively. Base on the method above, three order approximate solution for the problem of two dimensional reaction diffusion model is as follows,
where and are the functions which depend on the parameters.
Brusselator model is a classical dissipative structure model, which has been studied by many researchers. The conditions for Turing bifurcation of Brusselator model is as follows,
3.2. The Effective Range of h
Parameters have the ability to regulate and control the convergent region of series solution and the speed of convergence of the series solution. We can obtain an appropriate through the effective range about curve
of physical quantity and. The Figures 1-4 show the effective ranges of in different conditions. The effective ranges in Table 1 show that we will get different effective ranges under different condition of different parameters.
The effective area about <inline-formula><inline-graphic xlink:href="http://html.scirp.org/file/1-2340197x86.png" xlink:type="simple"/></inline-formula> of u, when<inline-formula><inline-graphic xlink:href="http://html.scirp.org/file/1-2340197x86.png" xlink:type="simple"/></inline-formula><inline-formula><inline-graphic xlink:href="http://html.scirp.org/file/1-2340197x87.png" xlink:type="simple"/></inline-formula>
The effective area about <inline-formula><inline-graphic xlink:href="http://html.scirp.org/file/1-2340197x89.png" xlink:type="simple"/></inline-formula> of v, when<inline-formula><inline-graphic xlink:href="http://html.scirp.org/file/1-2340197x89.png" xlink:type="simple"/></inline-formula><inline-formula><inline-graphic xlink:href="http://html.scirp.org/file/1-2340197x90.png" xlink:type="simple"/></inline-formula>
The effective area about <inline-formula><inline-graphic xlink:href="http://html.scirp.org/file/1-2340197x92.png" xlink:type="simple"/></inline-formula> of u, when<inline-formula><inline-graphic xlink:href="http://html.scirp.org/file/1-2340197x92.png" xlink:type="simple"/></inline-formula><inline-formula><inline-graphic xlink:href="http://html.scirp.org/file/1-2340197x93.png" xlink:type="simple"/></inline-formula>
The effective area about <inline-formula><inline-graphic xlink:href="http://html.scirp.org/file/1-2340197x95.png" xlink:type="simple"/></inline-formula> of v, when<inline-formula><inline-graphic xlink:href="http://html.scirp.org/file/1-2340197x95.png" xlink:type="simple"/></inline-formula><inline-formula><inline-graphic xlink:href="http://html.scirp.org/file/1-2340197x96.png" xlink:type="simple"/></inline-formula>
The effective area about <inline-formula><inline-graphic xlink:href="http://html.scirp.org/file/1-2340197x97.png" xlink:type="simple"/></inline-formula> under different parameters
Parameters
The effective area about of u
The effective area about of v
[−3, 1]
[−0.7, 0.5]
[−2.5, 1]
[−0.5, 0.5]
3.3. Simulation of Turing Patterns about Reaction-Diffusion
Parameter have the ability to regulate and control the convergent region of series solutions and the speed of convergence of the series solutions. The analysis of the basis function in the homotopy analysis method shows that when and are limited value, the homotopy series solutions are convergent. Choose the step and other parameters. In this case, as time goes by, the changes of Turing patterns structures show as Figures 5-12.
The patterns in space of u(t), when t = 0.1.
The patterns in space of v(t), when t = 0.1.
The patterns in space of u(t), when t = 0.5.
The patterns in space of v(t), when t = 0.5.
The patterns in space of u(t), when t = 1.
The patterns in space of v(t), when t = 1.
The patterns in space of u(t), when t = 10.
The patterns in space of v(t), when t = 10.
The parameters are constant and. In this case, as time goes by, the changes of Turing patterns structures show as Figures 13-20.
As time goes on, the final pattern structure will tend to a steady state. Turing patterns structure of homotopy series solutions are sensitive to the selection of initial guess solution and are affected by the wave number. In summary, in the range of parameter about Turing patterns, the system will appear striped patterns, point patterns and the coexistence of striped and point pattern with the time going on.
The patterns in space of u(t), when t = 0.1.
The patterns in space of v(t), when t = 0.1.
The patterns in space of u(t), when t = 0.5.
The patterns in space of v(t), when t = 0.5.
The patterns in space of u(t), when t = 1.
The patterns in space of v(t), when t = 1.
The patterns in space of u(t), when t = 10.
The patterns in space of v(t), when t = 10.
4. Conclusions
In this paper, a new method based on differential method and homotopy analysis method is used to solve the typical two-dimensional reaction diffusion model. Different shapes of Turing pattern can be obtained through Matlab mathematical software on experimental data simulation of the structure of the solution. And it is proved that the proposed method which to solving nonlinear reaction diffusion problems is feasible and effective. The new method not only reduces the dimension about differential in space, but also gets the analytical expression with physical parameters through the homotopy in time. It will facilitate the analysis of the influence of parameters variation on Turing pattern structure. The method which is differential in space and homotopy in time domain has enormous potential for solving definite solution of nonlinear partial differential, and it has great promotional value.
The innovation point of this article is making use of the new method to solve the two-dimensional reaction diffusion model. The new method is combining the differential method and homotopy analysis method.
Funding
This work is supported by the National Sciences Foundation of People’s Republic of China under Grant 1140011526.
Cite this paper
GendaiGu,HongxiaoPeng, (2015) Numerical Simulation of Reaction-Diffusion Systems of Turing Pattern Formation. International Journal of Modern Nonlinear Theory and Application,04,215-225. doi: 10.4236/ijmnta.2015.44016
ReferencesOuyang, Q. (2010) Nonlinear Science and Pattern Dynamics. Peking University Press, Beijing.Ouyang, Q. (2000) Reaction Diffusion System Pattern Dynamics. Shanghai Science and Technology Education Press, Shanghai.Liu P.P. ,et al. (2009)A Ratio Dependent Predator-Prey Model of Spatial Pattern Formation Research Mathematics in Practice and Theory 39, 114-119.Wang, Y., Cao, J.D., Sun, G.-Q. and Li, J. (2014) Effect of Time Delay on Pattern Dynamics in a Spatial Epidemic Model. Physica A: Statistical Mechanics and Its Applications, 412, 137-148. http://dx.doi.org/10.1016/j.physa.2014.06.038Parshad, R.D., Kumari, N., Kasimov, A.R. and Abderrahmane, H.A. (2013) Turing Patterns and Long-Time Behavior in a Three-Species Food-Chain Model. Mathematical Biosciences, 254, 83-102. http://dx.doi.org/10.1016/j.mbs.2014.06.007Liu, S.H. and Gu, Y.X. (2012) Coupled Reaction-Diffusion System in the Superlattice Pattern. Hebei University (Natural Science Edition), 32, 597-601.Du, Y.-K. and Xu, R. (2014) Pattern Formation in Two Classes of SIR Epidemic Models with Spatial Diffusion. Chinese Journal of Engineering Mathematics, 31, 454-462.Li, X.Z., Bai, Z.G., Li, Y., Zhao, K. and He, Y.F. (2013) Double Nonlinear Coupling Reaction-Diffusion Systems in Complex Turing Patterns. Chinese Journal of Physics, 62, 220503-1 -220503-7.