University of Twente Student Theses

Login

Graph Entropy on Fault Trees

Adam, E. (2024) Graph Entropy on Fault Trees.

[img] PDF
1MB
Abstract:A Fault Tree (FT) is a graphical model used in risk management to analyze systems. Graph entropy is a complexity measure that quantifies the structural information of graphical models. Given the limited availability of large, real world Fault Trees, creating more realistic randomly generated Fault Trees is important for developing and testing dedicated quantitative analysis tools for Fault Trees. This study explores the applicability of graph entropy in capturing the structural complexity of Fault Trees, by comparing randomly generated Fault Trees and real-life Fault Trees from the FFORT dataset, which includes a diverse range of real-world fault trees and other risk models sourced from scientific literature and industrial reports. It is intended as a starting point for enhancing the methods of random Fault Trees. To analyze this, the graph entropy of the in-degree and out-degree distributions was computed and compared across multiple graph sizes. The results in this paper show that real-world Fault Trees show higher entropy values as the graph scale increases, suggesting real-world graphs become structurally more predictable, whereas random Fault Trees show lower entropy values and these tend to stabilize at higher graph sizes. The findings display differences between structures and suggest that current methods of generating Fault Trees do not fully mimic the predictability of real-world Fault Trees. This research highlights the need for improved algorithms for generating Fault Trees.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science BSc (56964)
Link to this item:https://purl.utwente.nl/essays/100962
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page