Mutually catalytic branching in the plane: Finite measure states (Q1872307)

The purpose of the paper is to study the system of equations for \(x \in {\mathbb R}^2\): \[ \frac{\partial}{\partial t} X_t^i(x) = \frac{\sigma^2}{2} \Delta X_t^i(x) + \sqrt{ \gamma X_t^1(x) X_t^2(x)} \dot{W}_t^i (x) \tag{1} \] for \((t,x) \in {\mathbb R}_+ \times {\mathbb R}^2\), \(i= 1,2\), where \(\sigma > 0\) is the migration rate, \(\gamma > 0\) is the collision rate, and \(\dot{W}_t^i\) is a standard time-space white noise on \({\mathbb R}_+ \times {\mathbb R}^2\). This system is interactive in the sense that the branching rate of each type is proportional to the local density of the other type. The existence of this truly interacting system consisting of two types (where each type catalyzes the branching of the other) makes the situation in (1) quite different from the case without interaction. The one-dimensional case (i.e. the system (1) for \((t,x) \in\) \({\mathbb R}_+ \times {\mathbb R}\)) was already treated by \textit{D. A. Dawson} and \textit{E. A. Perkins} [Ann. Probab. 26, No.~3, 1088-1138 (1998; Zbl 0938.60042)]. However, existence of solutions to the system (1) in \({\mathbb R}^2\) seems to be counterintuitive, and the question to ask whether such an interacting system exists or not is highly nontrivial. The authors prove that, under the setting of sufficiently large diffusion rate \(\sigma^2\) compared with the branching rate \(\gamma\) (i.e. \(3 \sqrt{6} \pi C_0 \gamma < \sigma^2\) with \(C_0 \in\) \((3/20, 11/20)\)), the mutually catalytic model exists as the unique pair \({\mathbb X} = (X^1, X^2)\) of finite measure-valued processes via martingale formulation involving the collision local time \(L_X\) of the solutions with additional strong conditions on integrability of \({\mathbb X}\). Moreover, it is also shown that the processes have densities at fixed times which lie on disjoint sets and explode as they approach the interface of the two populations. As a matter of fact, the proof of uniqueness is given in a companion paper by \textit{D. A. Dawson, K. Fleischmann, L. T. Mytnik, E. A. Perkins} and \textit{J. Xiong} [Ann. Inst. Henri Poincaré, Probab. Stat. 39, No.~1, 135-191 (2003)]. And the integrability conditions to guarantee the uniqueness are verified in this paper as well. Lastly, they show the global extinction of one type in the long-term limit, which is an extension result of the one-dimensional version of Theorem 6.6 (p. 1125) proved in the first quoted paper. For some other related works on long time behaviors, see e.g. \textit{D. A. Dawson} and \textit{K. Fleischmann} [Stochastic Processes Appl. 71, No. 2, 241-257 (1997; Zbl 0942.60073)]. Although in this paper they are restricted themselves to the case of finite measure states as mutually catalytic branching model in the plane, the discussion on the case of infinite measure states can be found in the same authors' work [Electron. J. Probab. 7, No.~15, 1-61 (2002)].