Abstract
Phased mission systems (PMSs) involve a sequence of multiple tasks that must be carried out in consecutive, non-overlapping phases of operation. During each phase, the system is typically subject to different stresses and dependability requirements. Three major factors contribute to the difficulty in analysing PMSs: dynamics in system configuration, failure criteria, and component failure behaviour across different phases; s-dependence across the consecutive phases for a given component; and s-dependence among different components. A phase modular approach that integrates combinatorial solutions to static modules and Markov solutions to dynamic modules as appropriate has been recently proposed to address the above difficulty. Although the modular approach outperforms both the purely combinatorial solutions in modelling power and state-space based methods in computational efficiency, it involves costly path enumeration as well as joint phase module model generation and probability computation. This paper proposes improvements on the phase modular approach that can avoid the aforementioned costly operations and thus facilitate more efficient evaluation of hybrid PMSs. The proposed solutions are verified and illustrated through the analyses of an example phased mission system.