The scaling limit of the interface of the continuous-space symbiotic branching model

Stochastic partial differential equations (aspects of stochastic analysis) (60H15) Branching processes (Galton-Watson, birth-and-death, etc.) (60J80) Interacting random processes; statistical mechanics type models; percolation theory (60K35) Population dynamics (general) (92D25)

Abstract: The continuous-space symbiotic branching model describes the evolution of two interacting populations that can reproduce locally only in the simultaneous presence of each other. If started with complementary Heaviside initial conditions, the interface where both populations coexist remains compact. Together with a diffusive scaling property, this suggests the presence of an interesting scaling limit. Indeed, in the present paper, we show weak convergence of the diffusively rescaled populations as measure-valued processes in the Skorokhod, respectively the Meyer-Zheng, topology (for suitable parameter ranges). The limit can be characterized as the unique solution to a martingale problem and satisfies a "separation of types" property. This provides an important step toward an understanding of the scaling limit for the interface. As a corollary, we obtain an estimate on the moments of the width of an approximate interface.

Review text

Let \((u_t,v_t)\) be the nonnegative solution of the system NEWLINE\[NEWLINE\begin{aligned} {\partial\over\partial t} u_t(x) &= {\Delta\over 2} u_t(x)+ \sqrt{\gamma u_t(x)v_t(x)}\dot W^{(1)}_t(x),\\ {\partial\over\partial t} v_t(x) &= {\Delta\over 2} v_t(x)+ \sqrt{\gamma u_t(x)v_t(x)}\dot W^{(2)}_t(x),\end{aligned}NEWLINE\]NEWLINE with \((u_0,v_0)= (\mathbf{1}_{\mathbb{R}^-},\text\textbf{1}_{\mathbb{R}^+})\), \(\gamma>0\), and \((\dot W^{(1)},\dot W^{(2)})\) a pair of correlated standard Gaussian white noises on \(\mathbb{R}_+\times\mathbb{R}\) with correlation \(\rho\in[-1,1]\), see, e.g., [\textit{A. M. Etheridge} and \textit{K. Fleischmann}, Stochastic Processes Appl. 114, No. 1, 127--160 (2004; Zbl 1072.60086)]. Define the measure-valued processes NEWLINE\[NEWLINE\mu^{(n)}_t(dx):= u_{n^2t}(nx)\,dx,\qquad \nu^{(n)}_t(dx):= v_{n^2t}(nx)\,dx.NEWLINE\]NEWLINE If \(\rho\in(-1,-{1\over\sqrt{2}})\), then \((\mu^{(n)}_t, \nu^{(n)}_t)_{t\geq 0}\) is shown to converge, as \(n\to\infty\), with respect to the Skorokhod topology to the unique solution \((\mu_t, \nu^{(n)}_t)_{t\geq 0}\) of a martingale problem, satisfying \(\mu_t(dx)= \mu_t(x)\,dx\), \(\nu_t(dx)= \nu_t(x)\,dx\), \(\mu_t\nu_t= 0\). An explicit representation of the solution is not given, but a uniform estimate of the \(p\)-th moment, \(p<1\), of an approximation to the interface width \(|\sup\{x: u_t(x)> 0\}- \text{inf}\{x: v_t(x)>0\}|\) is derived, remarkably without the ``no-hole condition of \textit{C. Mueller} and \textit{R. Tribe} [Electron. J. Probab. 2, Paper 7, 27 p. (1997; Zbl 0890.60056)].

Recommendations

Cites work

Cited in

(8)

This page was built for publication: The scaling limit of the interface of the continuous-space symbiotic branching model

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q282485)