Reinforcement in Cooperative Games

Μπαρδής, Κωνσταντίνος; Bardis, Konstantinos

dc.contributor.author	Μπαρδής, Κωνσταντίνος	el
dc.contributor.author	Bardis, Konstantinos	en
dc.date.accessioned	2022-10-14T10:35:47Z
dc.date.available	2022-10-14T10:35:47Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/55942
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.23640
dc.description	Εθνικό Μετσόβιο Πολυτεχνείο--Μεταπτυχιακή Εργασία. Διεπιστημονικό-Διατμηματικό Πρόγραμμα Μεταπτυχιακών Σπουδών (Δ.Π.Μ.Σ.) “Επιστήμη Δεδομένων και Μηχανική Μάθηση"	el
dc.rights	Αναφορά Δημιουργού-Όχι Παράγωγα Έργα 3.0 Ελλάδα	*
dc.rights	Αναφορά Δημιουργού-Όχι Παράγωγα Έργα 3.0 Ελλάδα	*
dc.rights.uri	http://creativecommons.org/licenses/by-nd/3.0/gr/	*
dc.subject	Πολλαπλοί Πράκτορες	el
dc.subject	Ενισχυτική Μάθηση	el
dc.subject	Βαθιά Μάθηση	el
dc.subject	Νευρωνικά Δίκτυα	el
dc.subject	Βιντεοπαιχνίδια	el
dc.subject	Multi-Agent	en
dc.subject	Reinforcement Learning	en
dc.subject	Deep Learning	en
dc.subject	Neural Networks	en
dc.subject	Videogames	en
dc.title	Reinforcement in Cooperative Games	en
heal.type	masterThesis
heal.secondaryTitle	Deep Learning Approaches	en
heal.classification	Machine Learning	en
heal.classification	Reinforcement Learning	el
heal.classification	Deep Learning	el
heal.language	en
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2022-05-05
heal.abstract	A Multi-Agent system is a system that necessitates the coordination and interaction between several decision-making entities (Agents) to accomplish a given task they otherwise would not be able to. There is a growing need for algorithms tailored to this setting, since modern systems are relying more and more on decentralized cooperation between several agents, which poses several additional challenges over the more well studied single-agent setting, like the non-stationarity induced by other agents’ decisions, which behooves some important modifications to existing algorithms or the development of dedicated approaches from the ground up. This diploma thesis aims to first study the modern literature on MARL, explain the challenges and opportunities it affords, and then to utilize several of these algorithms in multi-agent settings using the libraries PettingZoo and RLLib in cooperative settings.	en
heal.advisorName	Κόλλιας, Στέφανος
heal.committeeMemberName	Κόλλιας, Στέφανος
heal.committeeMemberName	Σταφυλοπάτης, Ανδρέας-Γεώργιος
heal.committeeMemberName	Στάμου, Γιώργος
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Ηλεκτρολόγων Μηχανικών και Μηχανικών Υπολογιστών	el
heal.academicPublisherID	ntua
heal.numberOfPages	69
heal.fullTextAvailability	false