Development of a Parallel BAT and Its Applications in Binary-state Network Reliability Problems

TL;DR

This paper introduces a parallel version of the binary-addition-tree (BAT) algorithm, called PBAT, which leverages multithread architecture to efficiently solve medium-scale binary-state network reliability problems, overcoming previous scalability limitations.

Contribution

The paper develops a parallel BAT (PBAT) algorithm that significantly improves the scalability and efficiency of exact network reliability calculations for medium-sized networks.

Findings

PBAT outperforms existing implicit enumeration algorithms.

PBAT efficiently solves medium-scale network reliability problems.

Experimental results on 20 benchmarks validate the effectiveness of PBAT.

Abstract

Various networks are broadly and deeply applied in real-life applications. Reliability is the most important index for measuring the performance of all network types. Among the various algorithms, only implicit enumeration algorithms, such as depth-first-search, breadth-search-first, universal generating function methodology, binary-decision diagram, and binary-addition-tree algorithm (BAT), can be used to calculate the exact network reliability. However, implicit enumeration algorithms can only be used to solve small-scale network reliability problems. The BAT was recently proposed as a simple, fast, easy-to-code, and flexible make-to-fit exact-solution algorithm. Based on the experimental results, the BAT and its variants outperformed other implicit enumeration algorithms. Hence, to overcome the above-mentioned obstacle as a result of the size problem, a new parallel BAT (PBAT) was proposed to improve the BAT based on compute multithread architecture to calculate the binary-state network reliability problem, which is fundamental for all types of network reliability problems. From the analysis of the time complexity and experiments conducted on 20 benchmarks of binary-state network reliability problems, PBAT was able to efficiently solve medium-scale network reliability problems.