I found in the book Markov Chains by Revuz the following definition of a Markov chain. In the following $(X_n)_{n \in \mathbb{N}}$ is a sequence of random variables on a probability space $(\Omega,\mathcal{F},P_0)$ and the range is a measurable space $(E,\Sigma)$, and $\mathcal{F}_n^0 =\sigma(X_m , m \leq n)$

Definition.1 $(X_n)_{n \in \mathbb{N}}$ is a Markov chain if for every $B \in \Sigma$ we have $$P_0 [ X_m \in A \mid \mathcal{F}_n^0 ] = P_0[ X_m \in A \mid X_n] $$ for all $m\geq n$

For a transition probability $P : E \times \Sigma \to [0,1]$, i.e $x \to p(x,A)$ is measurable for all $A\in \Sigma$ and $A\to p(x,A)$ is a probability measure for all $x\in E$ he defines a family of transition probabilities $(P_n)_{n \geq 0}$ by $$ P_{n} (x,A) = \int_E p(y,A) \, P_{n-1}(x,dy) $$ and $$ P_0 := P.$$ Then he defines a homogeneous Markov chain by

Definition 2. $(X_n)_{n \in \mathbb{N}}$ is called a homogeneous Markov chain with respect to transition probability $P$ if $$ P_0[X_m \in A \mid \mathcal{F}_n^o] = P_{n-m}(x,\{X_m \in A\}).$$

I would like to know if $(X_n)$ is a Markov chain with $$ P_0[X_n \in B \mid X_0 \in A] = P_0[X_{n+m} \in B \mid X_m \in A] $$ for all $n,m\geq 0$ and $A,B \in \Sigma$, does this imply that one can find a transition probability $P$ such that $(X_n)$ is a homogeneous Markov chain with respect to $P$?


Thankfully, YES :)

You can very simply define the needed probability transition by :

\begin{equation} P : x, \mathcal{A} \mapsto P_0[X_1 \in \mathcal{A} | X_0=x] \end{equation}

You can then deduce from the definition of the family $(P_n)$ that :

\begin{equation} P_n ( x, \mathcal{A} ) = P_0[X_{n+1} \in \mathcal{A} | X_0=x] \end{equation}


Then verify that your Markov chain satisfies the condition to be homogeneous with respect to $(P_n)$, by using the definition of a Markov chain and the property.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.