Solving reliability problems in complex networks with approximated cuts and paths

Abstract

The paper solves reliability problems for the design of complex network without failure. The traditional approaches based on minimal paths and cuts require significant computational effort equals to NP-hard. The major difficulty lies in calculating the minimal cuts and paths to improve exact reliability bounds. Therefore, a neural network algorithm based on approximated paths and cuts to deal with this difficulty. The proposed approach is divided into two parts. The first part is used to approximate the computed minimal cuts and paths from two networks. The approach in other part improves the reliability bounds. The proposed approach has been tested on mesh network of 256 nodes and hyper-tree of 496 nodes. The performance of proposed approach is compared with PSO for the reliability bounds improvement in low failure probability.