A fast, general-purpose algorithm for reliability evaluation of distributed systems

A fast and general-purpose algorithm for reliability evaluation of distributed systems is developed. The algorithm is general in the sense that it is used to evaluate different network reliability measures, such as terminal, K-terminal, all-terminal, and degraded system reliability. Furthermore, it is used to evaluate the above measures in networks which are directed or undirected, as well as in networks with perfect or imperfect nodes. The options for numeric or symbolic and exact or approximate reliability computation are also available. The algorithm efficiently calculates reliability of distributed systems with size of practical interest and has comparative or much better speed than other reported algorithms. Also presented is a method for the very efficient handling of the case of imperfect nodes, an issue which is particularly important for distributed systems with unreliable processing nodes.<
>