3.3.2 How to find \(f_{ij}\)

This is the probability that chain will eventually reach state \(j\) given it starts in state \(i\)

\(i\in C(J)\)
\(j\in C(J)\)
\(f_{ij}=1\)
\(i\in C(J)\)
\(j\notin C(J)\)
\(f_{ij}=0\)
\(i\in T\)
\(j\) recurrent but not absorbent, hence in a closed set with other states
Use formula in page 5.5 lecture notes \(f_{ij}={\displaystyle \sum \limits _{k\in T}} p_{ik}f_{kj}+{\displaystyle \sum \limits _{k\in C\left ( J\right ) }} p_{ik}\) \(F=\left ( I-Q\right ) ^{-1}Z\) , hence just needs to find \(Z\)
\(z_{ij}=\)probability that transient state \(i\) will reach class that contains \(j\)
\(i\in T\)
\(j\) is an absorbent
\(f_{ij}=\left [ \left ( I-Q\right ) ^{-1}R\right ] _{i,j}\)
\(i\in T\)
\(j\in T\)
We know eventually \(p_{ij}=0\) for \(i,j\in Q\), but can we talk about \(f_{ij}\) here?