Large time asymptotics for the density of a branching Wiener process

Pál Révész,Jay Rosen,Zhan Shi

2004-09-27

Abstract: Given an R^d-valued supercritical branching Wiener process, let D(A,T) be the number of particles in a subset A of R^d at time T, (T=0,1,2,...). We provide a complete asymptotic expansion of D(A,T) as T goes to infinity, generalizing the work of X.Chen.

Probability

What problem does this paper attempt to address?

The problem that this paper attempts to solve is about the asymptotic behavior of the number of particles in a branching Wiener process over a long period of time. Specifically, the paper focuses on the complete asymptotic expansion of the number of particles $\psi(A, T)$ in a supercritical branching Wiener process in the high - dimensional space $\mathbb{R}^d$ and located within a certain set $A\subset\mathbb{R}^d$ as time $T\rightarrow\infty$.

### Main problem description:

1. **Model description**:
- At the initial moment, a particle starts from the origin and performs a Wiener process $W(t)$ in $\mathbb{R}^d$.
- When the particle reaches time $t = 1$, it dies and is replaced by $Y$ offspring particles, where $Y$ is a random variable satisfying:
\[
P\{Y=\ell\}=p_\ell,\quad\sum_{\ell = 0}^{\infty}\ell p_\ell=m<\infty,\quad\sum_{\ell = 0}^{\infty}(\ell - m)^2p_\ell=\sigma^2<\infty
\]
- Each offspring particle starts from the position where its ancestor died and repeats the above process. All Wiener processes and the number of offspring are independent.

2. **Research objective**:
- Given a supercritical branching Wiener process, define $\psi(A, T)$ as the number of particles located within the set $A\subset\mathbb{R}^d$ at time $T$.
- The goal of the paper is to provide the complete asymptotic expansion of $\psi(A, T)/m^T$ as $T\rightarrow\infty$.

3. **Existing results**:
- Chen (2001) has already proven that for any Borel set $A\subset\mathbb{R}^d$, as $T\rightarrow\infty$, there exist random variables $N_1$ and $N_2$ such that:
\[
\frac{(2\pi T)^{d/2}\psi(A, T)}{m^T}=N_0\int_A dx-\frac{1}{2T}\int_A(N_0\|x\|^2 - 2N_1\cdot x+N_2)dx+o(T^{- 1})
\]

4. **Main contributions of this paper**:
- It provides a more complete asymptotic expansion of $\psi(A, T)/m^T$, in the following form:
\[
\frac{(2\pi T)^{d/2}\psi(A, T)}{m^T}=\sum_{n = 0}^k(-T)^{-n}\sum_{|\alpha|=n}\frac{1}{\alpha!}\sum_{\beta\preceq2\alpha}\binom{2\alpha}{\beta}(-1)^{|\beta|}M_\beta(A)N_{2\alpha-\beta}+o(T^{-k})
\]
where $M_\alpha(A)=\int_A x^\alpha dx$, and $N_\alpha$ are certain random variables.

Through these studies, the paper aims to more accurately describe the long - time behavior of particle distribution in the branching Wiener process and provides a higher - order asymptotic expansion, so as to better understand and predict the particle concentration and other related characteristics in this process.