抄録
Markov's chains (MC) are well-know and widely applied in dependability and performability analysis of safety-critical systems, because of the flexible representation of system components dependencies and synchronization. There are few roadblocks for greater application of the MC: accounting the additional system components increases the model state-space and complicates analysis; the non-numerically sophisticated user may find it difficult to decide between the variety of numerical methods to determine the most suitable and accurate for their application. Thus obtaining the high accurate and trusted modeling results becomes a nontrivial task. In this paper, we present the metric-based approach for selection of the applicable solution approach, based on the analysis of MCs stiffness, decomposability, sparsity and fragmentedness. Using this selection procedure the modeler can provide the verification of earlier obtained results. The presented approach was implemented in utility MSMC, which supports the MC construction, metric-based analysis, recommendations shaping and model solution. The model can be exported to the well-known off-the-shelf mathematical packages for verification. The paper presents the case study of the industrial NPP I&C system, manufactured by RPC Radiy. The paper shows an application of metric-based approach and MSMC tool for dependability and safety analysis of RTS, and procedure of results verification.
