Abstract
Multistate phased-mission systems (MS-PMS) are multistate systems subject to multiple, consecutive, and nonoverlapping phases of operation. The challenges in analyzing MS-PMS reside in the dynamic system configuration, failure criteria, and component state transition behavior in different phases, as well as the s-dependence across different phases and among different states of a given component. Existing methods for the reliability analysis of MS-PMS are either based on monolithic Markov models that suffer from the well-known state explosion problem, or using a hierarchical strategy that can only handle ordered component states. This paper presents integrated modeling approaches for the reliability analysis of repairable MS-PMS with both ordered and unordered component states. The proposed methods integrate efficient decision diagram models for representing the system structure function and incorporating the unordered/ordered component states at the system level, and Markov models for describing dependence and transition behaviors at the component level. The application and advantages of the proposed approaches are illustrated through a case study in which the reliability for a sequence of tasks in a multistate distributed computing system is analyzed.