• PDF
• Cite
• Share
Article Contents  Article Contents

# Application of aggregation of variables methods to a class of two-time reaction-diffusion-chemotaxis models of spatially structured populations with constant diffusion

• * Corresponding author: anouarelharrak1@gmail.com
• The main goal of this paper is to adapt a class of complexity reduction methods called aggregation of variables methods to the construction of reduced models of two-time reaction-diffusion-chemotaxis models of spatially structured populations and to provide an error bound of the approximate dynamics. Aggregation of variables methods are general techniques that allow reducing the dimension of a mathematical dynamical system. Here we reduce a system of Partial Differential Equations to a simpler Ordinary Differential Equation system, provided that the evolution processes occur at two different time scales: a slow one for the demography and a fast one for migrations and chemotaxis, with a ratio $\varepsilon>0$ small enough. We give an approximation of the error between solutions of both original and reduced model for a generic function representing the demography. Finally, we provide an optimization of the error bound and validate numerically this result for a spatial inter-specific model with constant diffusion and population growth given by a logistic law in population dynamics.

Mathematics Subject Classification: Primary: 37M99, 37N25; Secondary: 35K55.

 Citation: • • Figure 1.  Plot of the numerical solution of the perturbed problem (11), (A) and (C), and its approximate solution using aggregation of variables methods, (B) and (D), at two times $t = 0.2$ and $t = 25$ for $\varepsilon = 1e-1$

Figure 2.  Plot of total population of the global model and its approximate solution using aggregation of variables methods for for $\varepsilon = 1e-1$. $K_T: = \int_{\Omega}K(x)dx$ stands for the total carrying capacity of the environment and $K^\ast$ for the new homogeneous one

Figure 3.  Plot of errors, $|N_{\varepsilon}(t)-N(t)|$, (A), and ${\Vert n_{\varepsilon}(.,t)-\lambda(.)N(t) \Vert_\infty}$ with $t\geq 1$, (B), with respect to time $t$ for different values of $\varepsilon$; $\varepsilon = 1e-1$, $\varepsilon = 1e-2$, $\varepsilon = 1e-3$, and $\varepsilon = 1e-4$

• ## Article Metrics  DownLoad:  Full-Size Img  PowerPoint