Calculation of node-pair reliability in large networks with unreliable nodes

A new efficient method that compensates for unreliable nodes in network reliability computations is presented. This method can be embedded in the modified Dotson algorithm or any algorithm that generates a symbolic reliability expression for networks with perfect nodes. Its cost increases linearly with the number of links, and the effect of unreliable nodes can be directly computed. This method supplants the Aggarwal method and other methods of compensating for unreliable nodes in the calculation of node-pair reliability. When combined with the modified Dotson algorithm, this method provides accurate reliability estimates for networks so large that the Theologou-Carlier algorithm cannot complete its computations in a reasonable amount of time. For such networks, the new method can be embedded in the modified Dotson algorithm to estimate both the node-pair reliability and the error in this estimate even if the algorithm is terminated before completion. >