recurrence in a Markov chain
Let be a stationary (http://planetmath.org/StationaryProcess) Markov chain![]()
and the state space
. Given and any non-negative integer , define a number as follows:
In other words, is the probability that the process first reaches state at time from state at time .
From the definition of , we see that the probability of the process reaching state within and including time from state at time is given by
As , we have the limiting probability of the process reaching eventually from the initial state of at , which we denote by :
Definitions. A state is said to be recurrent or persistent if , and transient otherwise.
Given a recurrent state , we can further classify it according to “how soon” the state returns after its initial appearance. Given , we can calculate the expected number of steps or transitions required to return to state by time . This expectation is given by
When , the above expression may or may not approach a limit. It is the expected number of transitions needed to return to state at all from the beginning. We denote this figure by :
Definitions. A recurrent state is said to be or strongly ergodic if , otherwise it is called null or weakly ergodic. If a stronly ergodic state is in addition aperiodic (http://planetmath.org/PeriodicityOfAMarkovChain), then it is said to be an ergodic state.
| Title | recurrence in a Markov chain |
| Canonical name | RecurrenceInAMarkovChain |
| Date of creation | 2013-03-22 16:24:43 |
| Last modified on | 2013-03-22 16:24:43 |
| Owner | CWoo (3771) |
| Last modified by | CWoo (3771) |
| Numerical id | 5 |
| Author | CWoo (3771) |
| Entry type | Definition |
| Classification | msc 60J10 |
| Synonym | null recurrent |
| Synonym | positive recurrent |
| Synonym | strongly ergodic |
| Synonym | weakly ergodic |
| Defines | recurrent state |
| Defines | persistent state |
| Defines | transient state |
| Defines | null state |
| Defines | positive state |
| Defines | ergodic state |