Convergence of stopped/unstopped Markov chains

87 Views Asked by At

Does weak convergence of stopped Markov chains implies weak convergence of the corresponding unstopped Markov chains in some particular case?

In my case:

Assumptions on the sequence of Markov chains:

$\forall n \in \mathbb{N}$

  • $\{X^n(t)\}_{t\geq 0 }\subset \frac{1}{n}\mathbb{N}^d\setminus\{\mathbf{0}\}$ Markov chain
  • $\tau_n=\inf\{t:X_i^n(t)= \frac{1}{n} \text{ for some } i\}$ is a finite stopping time
  • if $s\geq t$ then $||X^n(s)||_1\leq ||X^n(t)||_1$
  • $X^n(0)\to \mathbf{x}_0\in \mathbb{R}_{>0}^d$

Assumptions on the limiting process:

  • $\{X(t)\}_{t\geq 0 } \subset \mathbb{R}_{\geq 0}^d$ strongly continuous Markov process
  • $X(0)=\mathbf{x}_0$
  • $\forall t\geq ||\mathbf{x}_0||_1, X(t)=\mathbf{0}$.
  • $\inf\{t:X_i(t)= 0 \text{ for some } i\}=||\mathbf{x}_0||_1$

Weak convergence:

  • We know that $\{X^n(t\wedge \tau_n)\}_{t\geq 0}$ converges weakly to $X$.

I am trying to prove that $X^n$ should also converge weakly to $X$, because, heuristically, it seems that this is implied by $\tau_n\to ||\mathbf{x}_0||_1$, $X^n(\tau_n)\to X(||\mathbf{x}_0||_1)=\mathbf{0}$ and for $t\geq \tau_n$, $||X^n(t) ||_1\leq ||X^n(\tau_n)||_1\to 0$.

Is my intuition correct? Is there any classical result that could help me here?