QB-II for Evaluating the Reliability of Binary-State Networks

TL;DR

The paper introduces QB-II, an improved algorithm for evaluating the reliability of binary-state networks, enhancing efficiency over existing methods through innovative graph partitioning and matrix convolution techniques.

Contribution

It presents QB-II, a novel algorithm that outperforms previous algorithms in assessing binary-state network reliability using advanced graph partitioning and matrix operations.

Findings

QB-II outperforms many existing algorithms in benchmark tests.

The algorithm efficiently computes network reliability using shortest minimum cuts.

Validation on twenty benchmark problems confirms its effectiveness.

Abstract

Current real-life applications of various networks such as utility (gas, water, electric, 4G/5G) networks, the Internet of Things, social networks, and supply chains. Reliability is one of the most popular tools for evaluating network performance. The fundamental structure of these networks is a binary state network. Distinctive methods have been proposed to efficiently assess binary-state network reliability. A new algorithm called QB-II (quick binary-addition tree algorithm II) is proposed to improve the efficiency of quick BAT, which is based on BAT and outperforms many algorithms. The proposed QB-II implements the shortest minimum cuts (MCs) to separate the entire BAT into main-BAT and sub-BATs, and the source-target matrix convolution products to connect these subgraphs intelligently to improve the efficiency. Twenty benchmark problems were used to validate the performance of the QB-II.