Strong law of large numbers for supercritical superprocesses under second moment condition

TL;DR

This paper establishes a strong law of large numbers for supercritical superprocesses on a metric space, showing almost sure convergence of scaled process functionals under second moment conditions.

Contribution

It proves a strong law of large numbers for supercritical superprocesses with general spatial motion and branching mechanisms, extending previous results under second moment conditions.

Findings

Almost sure convergence of scaled superprocess functionals.

Identification of the limit involving eigenfunctions and martingale limits.

Applicability to a broad class of functions and initial measures.

Abstract

Suppose that $X=\{X_t, t\ge 0\}$ is a supercritical superprocess on a locally compact separable metric space $(E, m)$. Suppose that the spatial motion of $X$ is a Hunt process satisfying certain conditions and that the branching mechanism is of the form $$ \psi(x,\lambda)=-a(x)\lambda+b(x)\lambda^2+\int_{(0,+\infty)}(e^{-\lambda y}-1+\lambda y)n(x,dy), \quad x\in E, \quad\lambda> 0, $$ where $a\in \mathcal{B}_b(E)$, $b\in \mathcal{B}_b^+(E)$ and $n$ is a kernel from $E$ to $(0,\infty)$ satisfying $$ \sup_{x\in E}\int_0^\infty y^2 n(x,dy)<\infty. $$ Put $T_tf(x)=\mathbb{P}_{\delta_x}< f,X_t>$. Let $\lambda_0>0$ be the largest eigenvalue of the generator $L$ of $T_t$, and $\phi_0$ and $\hat{\phi}_0$ be the eigenfunctions of $L$ and $\hat{L}$ (the dural of $L$) respectively associated with $\lambda_0$. Under some conditions on the spatial motion and the $\phi_0$-transformed semigroup of $T_t$, we prove that for a large class of suitable functions $f$, we have $$ \lim_{t\rightarrow\infty}e^{-\lambda_0 t}< f, X_t> = W_\infty\int_E\hat{\phi}_0(y)f(y)m(dy),\quad \mathbb{P}_{\mu}{-a.s.}, $$ for any finite initial measure $\mu$ on $E$ with compact support, where $W_\infty$ is the martingale limit defined by $W_\infty:=\lim_{t\to\infty}e^{-\lambda_0t}< \phi_0, X_t>$. Moreover, the exceptional set in the above limit does not depend on the initial measure $\mu$ and the function $f$.