Synthesizing Resource-Shielded Policies for Partially Observable Markov Decision Processes – Bc. Šimon Brlej
Bc. Šimon Brlej
Bachelor's thesis
Synthesizing Resource-Shielded Policies for Partially Observable Markov Decision Processes
Synthesizing Resource-Shielded Policies for Partially Observable Markov Decision Processes
Abstract:
Čiastočne pozorovateľné Markovove Rozhodovacie Procesy (POMDP) s obmedzenými zdrojmi umožňujú modelovanie prostredia, kde agent potrebuje sledovať zmenšujúci sa zdroj s neistými pozorovaniami jeho pozície v danom prostredí. Cieľom tejto práce bolo vytvoriť nástroj, v ktorom je implementovaný nový algoritmus pre optimalizáciu bezpečného dosahovania cieľov v cieľovo zameraných POMDP s obmedzenými zdrojmi …moreAbstract:
Partially Observable Markov Decision Processes with resource constraints allow modeling of an environment where the agent needs to keep track of a diminishing resource under uncertain observations of its position in the environment. The goal of this thesis was to create a tool implementing a new algorithm for optimizing safe goal-reachability in goal-oriented resource-constrained POMDPs by combining …more
Language used: English
Date on which the thesis was submitted / produced: 19. 5. 2022
Identifier:
https://is.muni.cz/th/hvby2/
Thesis defence
- Date of defence: 30. 6. 2022
- Supervisor: RNDr. Petr Novotný, Ph.D.
- Reader: RNDr. Vít Musil, Ph.D.
Citation record
ISO 690-compliant citation record:
BRLEJ, Šimon. \textit{Synthesizing Resource-Shielded Policies for Partially Observable Markov Decision Processes}. Online. Bachelor's thesis. Brno: Masaryk University, Faculty of Informatics. 2022. Available from: https://theses.cz/id/zeqh1h/.
Full text of thesis
Contents of on-line thesis archive
Published in Theses:- světu
Other ways of accessing the text
Institution archiving the thesis and making it accessible: Masarykova univerzita, Fakulta informatikyMasaryk University
Faculty of InformaticsBachelor programme / field:
Informatics / Artificial Intelligence and Natural Language Processing