conditions on the matrix s.t. the matrix is not known explicitly although the stationary distribution is known.

328 Views Asked by At

We know that if the transition probability matrix of a Markov chain is doubly stochastic (column sum also 1) then the stationary distribution is known ($\frac{1}{\text{no. of states}}$) although the exact matrix is not known.

I am interested in knowing similar conditions on the matrix s.t. the matrix is not known explicitly although the stationary distribution is known.

1

There are 1 best solutions below

5
On

Let $P$ be a $n\times n$ (row-)stochastic matrix such that $\pi P = \pi$ where $\pi_i=\frac1n$, $1\leqslant i\leqslant n$. Then for each column $j$, we have $$\sum_{i=1}^n \pi_i P_{ij} = \pi_j, $$ hence $$\sum_{i=1}^n \frac1n P_{ij} = \frac1n, $$ and multiplying by $n$ we see that the columns sum to $1$. So the stationary distribution is uniform if and only if $P$ is doubly stochastic.