MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

This is a more specific version of a question I asked before without much luck. I believe this should be standard perturbation theory, but looking at Kato's book has not helped. Any references would be appreciated!

Let $X_n$ be a Markov chain on a compact metric space $(\Omega,d)$ with generator $\mathcal{L} \colon \mathcal{B}(\Omega) \to \mathcal{B}(\Omega)$, where $\mathcal{B}(\Omega)$ is the Banach space

$\mathcal{B}(\Omega) = \{ f \colon \Omega \to \mathbb{R} : \|f\|_{Lip}+ \|f\|_\infty < \infty\}$. (It doesn't really matter what the Banach space is)

Suppose the chain has a unique invariant measure $\mathbb{Q}$. Consider a perturbed Markov chain $X_n^\epsilon$ with Markov operator $\mathcal{L}_\epsilon = \mathcal{L} + \epsilon \Delta$, where $\Delta({1}) \equiv 0$ and $\|{\Delta}\| \leq 1$.

Suppose that for $\epsilon$ small enough, $\mathcal{L}_\epsilon$ has a unique invariant measure $\mathbb{Q}_\epsilon$.

  • Given that $\mathcal{L}_\epsilon$ is a compact operator from $\mathcal{B}(\Omega)$ to itself for $\epsilon$ small enough, with spectral gap independent of $\epsilon$, then for any test function $g\in\mathcal{B}(\Omega)$, is

    $\epsilon \mapsto \mathbb{Q}_\epsilon(g)$

    a real analytic function of $\epsilon$ around $0$?

Recognizing that the invariant measure is an eigenfunction of the adjoint operator has not really led to anything fruitful, though I'm by no means an expert in the field. Any help would be much appreciated!

share|cite|improve this question
The assumption that the generator is a compact operator appears quite restrictive to me. What example do you have in mind? – Wolfgang Loehr Sep 13 '12 at 9:14
Yes, it is restrictive, but I'm starting with a simpler model first to understand the techniques. I'm looking at a certain "environment viewed from the particle" process for a certain Random Walk in a Random Environment. When there's a strong enough drift, the operator of this process turns out to be compact in this restricted Banach space of functions. – Jeremy Voltz Sep 14 '12 at 17:23

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.