Questions: Probabilistic Model Checking

4 questions to test your understanding

Score: 0 / 4
Question 1 Short Answer

A Markov chain models a system where every transition has a fixed probability. An MDP (Markov decision process) additionally has nondeterministic choices. Why is the MDP model more expressive?

Think about your answer, then reveal below.
Question 2 Short Answer

In probabilistic temporal logic, the formula P(φ ≥ 0.95) means 'the probability of φ is at least 95%.' For a Markov chain, this is evaluated as: compute the probability of φ, check if it is ≥ 0.95. For an MDP, what does this formula mean?

Think about your answer, then reveal below.
Question 3 Multiple Choice

Rewards in probabilistic model checking allow reasoning about expected values beyond just probabilities. A system might have a property: 'the expected time to recover from failure is at most 10 seconds.' How would you verify this?

AAssign a time cost to each transition and use model checking to compute the expected cost (sum of costs on all paths, weighted by path probability). Check if this expected value is ≤ 10
BSimulate the system 1000 times and average the recovery times
CFormally verify the system cannot fail
DMeasure the system on a test run
Question 4 Short Answer

Probabilistic model checking requires solving systems of equations. For a Markov chain computing the probability of reaching a goal state, the equations are linear (probability of reaching goal from state s = sum over next states of (transition probability * probability from next state)). Why does this enable efficient checking?

Think about your answer, then reveal below.