Can finite state machines with conditional transitions be expressed as Markov chains?

Question

I'd be curious to know whether finite state machines that have conditional transitions can be expressed as Markov chains? If they can't, what would be a good counterexample?

Answer 1

A Markov Chain is a "recording" of a state machine where you have probablities between state changes. A UML state machine does not directly have these probablities. State transitions only have guards (and some other attributes). So in order to create a Markov Chain you need a profile which defines state transitions that also offer probabilities. Maybe (depends on your use case) you can simply use the guards to express these probabilities.