Analysis of Multi-State Systems using multi-valued decision diagrams

Abstract

A distinct characteristic of a multi-state system (MSS) is that the system and/or its components may exhibit multiple performance levels (or states) varying from perfect operation to complete failure. A MSS can model dependencies such as shared loads, performance degradation, imperfect fault coverage, standby redundancy, limited repair resources, and common-cause failures. The non-binary state property of a MSS and its components makes the analysis of such a system difficult. This paper proposes efficient algorithms for analyzing a MSS using multi-valued decision diagrams. Various reliability, availability, and performability measures are considered. The application and advantages of the proposed algorithms are demonstrated through an example.