Definition 4: For a time-homogeneous Markov chain,
which means that the Markov chain can be described by
a single, time-independent matrix P. Then the stationary
distribution Jr = (Jrl' Jr2, ... , Jr n) exists if the solution of the
equation pJr = Jr subject to L]=l Jrj = 1 exists.
We remark that if the steady-state probability distribution
of a PBN exists then it must be the stationary probability
distribution but not vice versa. Thus if we denote Jr as the
stationary distribution and Jr semi as the stationary distribution
regarding to probability structure matrix. It is natural that we
can define Jr semi as follows:
Definition 5: Noted that the PBN with Asemi as its probability
transition matrix is a Markov chain, so we define ( 1 2 n )T Jrs emi h . d· · = Jr semi' Jr semi' ... , Jr semi as t e statIOnary Istnbution,
which can be given by