EEL6507sp09L06
From BoykinWiki
Contents |
EEL6507 Spring 2009, Lecture 06, Wednesday 2008/01/21 (Notes created by Kyungyong Lee)
Basic Definitions for Markov Chain
- Transient state: fj < 1, where
- Recurrent state: fj = 1
- Recurrent null:
, where
= Average time to return to Ej
- Recurrent non-null:
- Recurrent null:
- Period: GCD of set {n:
}
- To be periodic, the period value should be bigger than 1.(Remember that if period value is 1, it is said to be aperiodic.)
- Ergodic state: A state Ej is said to be ergodic if it is aperiodic, and recurrent non-null.
- Ergodic markov chain: Every state is ergodic.
Basic Theorems for Markov Chain
- Theorem 1: The states of an irreducible Markov chain are either all transient or all recurrent nonnull or all recurrent null. If periodic, then all states have the same period value.
- Theorem 2: In an irreducible and aperiodic homogeneous(not time dependent) Markov chain, the limiting probabilities
always exists and are independent of the initial state probability distribution.
- (a) πj = 0 if all states are transient or recurrent null.
- (b)
, if all states are recurrent non null.
Getting Transient State Probabilities
- We can get Pn using z-transform by following steps,
Define
Hence,
So,
We know that
, and
Finally, we can derive
