Stochastic Modeling and Analysis of Phased-Mission Systems Dependability

SUMMARY & CONCLUSIONSThis paper presents a review of the stochastic modeling approaches in the literature for representing the phase-dependent behavior and quantitatively analyzing the dependability attributes of class of systems known as phased-mission systems (PMS, hereafter).The operational life (or mission) of a PMS can be split into a sequence of phases, each characterized by a specific workload, configuration, and operating conditions. Several dependability critical systems, such as aircraft control systems, missile flight systems and space probe navigation systems, exhibit phased-mission structures. Hence, their dependability modeling and evaluation is of utter relevance. However, PMSs offer several challenges to the modeling and quantitative analysis of dependability attributes, as their dynamic aspects can easily stretch the ability of classical modeling approaches. Because of it, researchers and practitioners in the field of dependability modelling and evaluation have devoted substantial effort to the definition of approaches that can accommodate phase-dependent behaviors of systems.The inherent complexity of PMSs has been managed resorting to combinatorial modeling methods, such as fault-trees, reliability diagrams and binary decision diagrams, as well as state-based representations such as Markov chains and Petri nets. Distinct approaches have been applied to represent the randomness of failure/repair event occurrence times, the possible uncertainty associated with phase ordering and duration, and the dependence that may exist between the residual lifetimes of system components in subsequent phases.As a result, a rich repertoire of proposed techniques is now available, which received contributions from many different research groups and integrated diverse approaches combining them in original ways. In the paper, we will explore these approaches and techniques, describing the challenges that drove the study of this class of complex systems and the outcomes of this yet unfinished research effort.