Importance analysis with Markov chains

In this paper, the authors introduce novel techniques for computing importance measures in state space dependability models. Specifically, reward functions in a Markov reward model (MRM) are utilized for this purpose, in contrast to the common method of computing importance measures through combinatorial models and structure functions. The advantage of bringing these measures in the context of MRMs is that the mapping extends the applicability of these substantial results of reliability engineering, previously considered only associated with fault trees and other combinatorial modeling techniques. As a consequence, software packages that allows the automatic description of MRMs can easily compute the importance measures under this new circumstance.