Network reliability analysis: Part I

Abstract

Abstract This paper considers networks with randomly failing links and nodes. In Part 1, nodes are assumed to be perfectly reliable. A combinatorial analysis is given when all links have equal reliabilities. Two general simulation methods are described. The first is particularly useful if a wide range of failure probabilities is to be considered. The second combines a combinatorial analysis with stratified sampling to yield major computational savings. Later parts will describe generalizations, decomposition methods for large networks and applications to computer network analysis.