0
$\begingroup$

Show that if $(X_n)_{n\geq 0}$ is a finite state irreducible Markov chain with transition matrix $P$, then a Markov chain with transition matrix $Q= \frac{1}{2}(I+P)$ is irreducible and aperiodic. Moreover, show that the two chains have the same invariant distribution.

I managed to show that the markov chain with transition matrix $Q$ is irreducible and aperiodic, but I'm stuck at showing they have the same invariant distribution.

I've played around a bit with that $\pi_1P=\pi_1$ and $\pi_2( \frac{1}{2}(I+P))=\pi_2$, where $\pi_1$ and $\pi_2$ are the invariant distributions. But I'm not really getting anywhere.

$\endgroup$
  • 1
    $\begingroup$ $\pi_2( \frac{1}{2}(I+P))=\pi_2 \implies \pi_2 P = \pi_2$, since $P$ is irreductible and aperiodic, the stationnary distribution is unique. $\endgroup$ – nicomezi May 28 '19 at 13:59
1
$\begingroup$

For $\pi_1$, $$ \begin{aligned} \pi_1P&=\pi_1 \end{aligned} $$

For $\pi_2$, $$ \begin{aligned} \pi_2[\frac{1}{2}(I+P)]&=\pi_2 \\ \pi_2[\frac{1}{2}(I+P)]-\pi_2&=0 \\ \pi_2[\frac{1}{2}(I+P)-I]&=0 \\ \pi_2[\frac{1}{2}(P-I)]&=0\\ \frac{1}{2}\pi_2(P-I)&=0 \\ \pi_2(P-I)&=0\\ \pi_2P-\pi_2&=0\\ \pi_2P&=\pi_2 \end{aligned} $$

I think this can answer your question.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.