Improved MDD Algorithm for Mission Reliability Estimation of an Escort Formation

An escort formation is a phased mission system of systems (PMSoS), which is composed of multiple ships with different functions. The configuration of the formation and success criteria for a mission may vary in different phases. Reliability estimation for PMSoS is complicated due to the strong phase dependence of multiple systems. This paper proposes an improved multiple-valued decision diagram (MDD) algorithm to perform reliability estimation of a nonrepairable escort formation. First, a phased fault tree is established to describe the failure mode of an escort formation throughout a mission, which is simplified according to the common failure basic mission (module) (CFBM). Bottom events are sorted based on the CFBM, and the case method is adopted to generate an MDD from the simplified fault tree model. On this basis, the MDD method is adopted to estimate mission reliability. The performance of the improved MDD method is compared with that of a binary decision diagram (BDD) method and a general MDD method. The results show that the improved MDD method can offer lower computational complexity as well as a simpler model construction over the BDD method and general MDD method. A case study of an escort formation PMSoS is analyzed to illustrate the proposed MDD method, and the sensitivity and composite importance measure (CIM) of each system are evaluated.