Next: Reversibility; Estimates for the
Up: Ergodicity and Stationarity
Previous: Stationary Initial Distributions
  Contents
Direct and Iterative Computation Methods
First we show how the stationary initial distribution
of the
Markov chain
can be computed based on methods from
linear algebra in case the transition matrix
does not
exhibit a particularly nice structure (but is quasi-positive) and
if the number
of states is reasonably small.
- Proof
-
- In order to prove that the matrix
is invertible
we show that the only solution of the equation
 |
(74) |
is given by
.
- As
satisfies the equation
we
obtain
 |
(75) |
- Thus (74) implies
i.e.
 |
(76) |
- On the other hand, clearly
and hence
as a consequence of (76)
and |
(77) |
- Taking into account (74) this implies
and, equivalently,
.
- Thus, we also have
for all
.
- Furthermore, Theorem 2.4 implies
,
- Thus, the matrix
is invertible.
- Finally, (75) implies
and, equivalently,
- Remarks
-
- Proof
-
- Proof
-
- Remarks
-
Next: Reversibility; Estimates for the
Up: Ergodicity and Stationarity
Previous: Stationary Initial Distributions
  Contents
Ursa Pantle
2006-07-20