Sensitivity computation of reliability Markov models using the uniformized power method

A new approach for computing the sensitivity of fault-tolerant systems modeled by a Markov process, as the derivative of the transient solution with respect to a system parameter, is proposed. We compare this approach with respect to the Standard Uniformization technique from a computation time point of view. This new method is faster than the Standard Uniformization method if the mission time is long and the state space is not too large.