Javascript must be enabled for the correct page display

A Centrality-Based Security Game for Asynchronous Multi-Hop Networks

Brinkhorst, R.J. (2016) A Centrality-Based Security Game for Asynchronous Multi-Hop Networks. Master's Thesis / Essay, Industrial Engineering and Management.

[img] Text
Master_IEM__Research_Project_2_1.pdf - Published Version
Restricted to RUG campus only

Download (621kB)
[img] Text
Toestemming.pdf - Other
Restricted to Backend only

Download (565kB)


This thesis, part of the University of Groningen’s Industrial Engineering and Management master degree, studies the performance of a distributed saddle-point equilibrium strategy algorithm on asynchronous multi-hop networks. The proposed algorithm computes the saddle-point equilibrium strategies for two-player zero-sum games. The players are an attacker that tries to minimize network performance and a defender, representing the asynchronous network nodes, that wants to maximize network performance. The algorithm is designed for the defender which has to calculate the saddle-point defense strategy distributively. The change in network performance caused by an attack corresponds to one of several node centrality measures. The algorithm is able to compute saddle-point equilibrium strategies for single node attack and single node defense games. For single node attack and multiple node defense games the algorithm requires more time to compute the strategies due to more complex computations. Simulations on numerous random networks are performed where the influence on algorithm performance by various network characteristics such as network size, centrality measure and asynchronism measure, are analyzed. The simulation results show that the algorithm converges to the saddle-point equilibrium strategies within reasonable time.

Item Type: Thesis (Master's Thesis / Essay)
Degree programme: Industrial Engineering and Management
Thesis type: Master's Thesis / Essay
Language: English
Date Deposited: 15 Feb 2018 08:24
Last Modified: 15 Feb 2018 08:24

Actions (login required)

View Item View Item