Treniranje agenata za provođenje napada u simulatoru CCS korištenjem okruženja CyberBatleSim

master's thesis

• Title (english): Training agents for attack execution in CCS simulator using CyberBatleSim environment
• Author: Luka Braut
• Mentor: Stjepan Groš (mentor)
• Committee member: Stjepan Groš (predsjednik povjerenstva)
• Committee member: Adrian Satja Kurdija (član povjerenstva)
• Committee member: Leonardo Jelenković (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2024-07-01, Croatia

Abstract

Ovaj rad istražuje treniranje agenata za provođenje napada u Cyber Conflict Simulatoru
(CCS) korištenjem okruženja CyberBattleSim. Započinje s pružanjem pregleda podržanog
učenja, uključujući osnovne koncepte, algoritme i njihovu primjenu u kibernetičkoj
sigurnosti. Razmatraju se arhitektura, komponente te postupak postavljanja i konfiguracije
simulacija u CyberBattleSimu. Poseban naglasak stavljen je na razvoj alata za preslikavanje
okruženja iz CCS – a u CyberBattleSim, ističući ključne razlike između ta dva sustava te
tehničke detalje implementacije. Na kraju, dan je pregled treniranja autonomnih agenata u
CyberBattleSimu, uključujući odabir i konfiguraciju agenata, metodologiju treniranja te
analizu njihovog ponašanja s posebnim naglaskom na ograničenja i potencijalna
poboljšanja.

Abstract (english)

This thesis investigates the training of agents for conducting attacks in the Cyber
Conflict Simulator (CCS) using the CyberBattleSim environment. It begins with an
overview of reinforcement learning, including basic concepts, algorithms, and their
application in cybersecurity. The architecture, components, and the process of setting up
and configuring simulations in CyberBattleSim are examined. Special emphasis is placed
on the development of a tool for mapping environments from CCS to CyberBattleSim,
highlighting the key differences between the two systems and the technical details of the
implementation. Finally, a review is given of training autonomous agents in
CyberBattleSim, including the selection and configuration of agents, training methodology,
and the analysis of their behavior with a focus on limitations and potential improvements.

• Keywords: Podržano učenje; CyberBattleSim; Cyber Conflict Simulator; Automatizacija napada
• Keywords (english): Reinforcement Learning; CyberBattleSim; Cyber Conflict Simulator; Attack Automation
• Language: croatian
• URN:NBN: urn:nbn:hr:168:608401
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2024-10-12 20:11:21