Nonlocal generalized models of predator-prey systems

From MaRDI portal
Publication:1948823



Abstract: The method of generalized modeling has been applied successfully in many different contexts, particularly in ecology and systems biology. It can be used to analyze the stability and bifurcations of steady-state solutions. Although many dynamical systems in mathematical biology exhibit steady-state behaviour one also wants to understand nonlocal dynamics beyond equilibrium points. In this paper we analyze predator-prey dynamical systems and extend the method of generalized models to periodic solutions. First, we adapt the equilibrium generalized modeling approach and compute the unique Floquet multiplier of the periodic solution which depends upon so-called generalized elasticity and scale functions. We prove that these functions also have to satisfy a flow on parameter (or moduli) space. Then we use Fourier analysis to provide computable conditions for stability and the moduli space flow. The final stability analysis reduces to two discrete convolutions which can be interpreted to understand when the predator-prey system is stable and what factors enhance or prohibit stable oscillatory behaviour. Finally, we provide a sampling algorithm for parameter space based on nonlinear optimization and the Fast Fourier Transform which enables us to gain a statistical understanding of the stability properties of periodic predator-prey dynamics.


Consider a predator-prey model of the form \[ \begin{aligned} X'&=S(X)-G(X,Y), \\ Y'&=G(X,Y)-M(Y),\end{aligned} \] where \(S,G,M\) are smooth functions. The method of generalized modeling is as follows. Assume that the system has at least one positive equilibrium \((X^*,Y^*)\). Using the transformation \(x=X/X^*\), \(y=Y/Y^*\), the equilibrium is moved to \((x^*,y^*)=(1,1)\) and system (1) can be rewritten as \[ \begin{aligned} x'&=\beta_1[s(x)-g(x,y)], \\ y'&=\beta_2[g(x,y)-m(y)] \end{aligned} \] with appropriately defined constants \(\beta_1,\beta_2\) and functions \(s,g,m\). Here, \(\beta_1,\beta_2\) are the so-called scale parameters. The Jacobian of the system at \((1,1)\) is associated with the so-called elasticities defined by \[ s_x=\partial_x(s(x))|_{x=1},\;g_x=\partial_x(g(x,y))|_{(x,y)=(1,1)}, \] \[ g_y=\partial_y(g(x,y))|_{(x,y)=(1,1)},\;m_y=\partial_y(m(y))|_{y=1}. \] The scale parameters and elasticities are termed generalized parameters. Stability analysis, bifurcation analysis, and other methods can be applied to the model in the generalized parameter space and obtain results that are valid for a wide class of models. In this paper, the method of generalized modeling is extended to systems with periodic solutions. Let \(\gamma(t)=(\gamma_1(t),\gamma_2(t))\) be a positive periodic orbit of system (1) with period \(T\). Similarly, using the transformation \(x=X/\gamma_1\), \(y=Y/\gamma_2\), the periodic orbit is reduced to the point \((1,1)\). System (1) now takes the form \[ \begin{aligned} x'&=\beta_s[s(x)-x]-\beta_1[g(x,y)-x], \\ y'&=\beta_2[g(x,y)-y]-\beta_m[m(y)-y]. \end{aligned} \] Then different methods, Floquet multipliers, Fourier series and statistical sampling, are used to obtain results on the stability of periodic solutions.





Describes a project that uses

Uses Software






This page was built for publication: Nonlocal generalized models of predator-prey systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1948823)