We have defined markov processes (in continuous time) as a collection of random variables $(X(t))_{t\in \mathbb{R_+}}$ such that in particular we have the property :

$P(X(t+s)=j|X(u), 0\leq u \leq t)=P(X(t+s)=j|X(t))$

So my question is : is this really the conditional probability on a random variable, $P(A|X)$ (I know this concept exists). Or does it really mean

$P(X(t+s)=j|X(t)=i, X(u)=x(u), 0\leq u < t)=P(X(t+s)=j|X(t)=i)$?


The definition in Grimmett & Stirzaker's Probability and Random Process states:

The process $X$ is called a (continuou-time) Markov chain if it satisfies the Markov property: $$P(X(t_n)=j|X(t_1)=i_1,\cdots, X(t_{n-1})=i_{n-1})= P(X(t_n)=j|X(t_{n-1})=t_{n-1})$$ for all $j,i_1,\cdots, i_{n-1} \in S$ and any sequence $t_1<t_2<\cdots <t_n$ of times.

I think this definition will make your understanding more clear.


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.