Steady State Probabilities

Do rij(n) converge to some πj (independent of i) ? where πj denotes the steady state probability of occupancy of state j, or P(Xn=j) for large n.

Yes if,

  • recurrent states are all in a single class
  • single recurrent class is not periodic (otherwise oscillations are possible)

Assuming yes, rij(n)=krik(n1)pkjlimnrij(n)=limnkrik(n1)pkjπj=kπkpkjbalance equationsand,iπi=1frequency of transitions kj=πkpkjin one stepfrequency of transitions into j=kπkpkjinflux from all connected states

The pij sum up to 1 and form a probability distribution called the stationary distribution of the chain (because if the initial distribution P(X0=j)=πj, the occupancy distribution of the states is constant for all steps and can be verified using total probability theorem on any of the nodes).

In the steady state,

  • πj=0 for transient states
  • πj>0 for recurrent states (note that any state that is absorbing is actually recurrent since its only connected to itself and hence accessible to itself from itself)