Stammdaten

A Novel Adaptive Weight Selection Algorithm for Multi-Objective Multi-Agent Reinforcement Learning
Untertitel:
Kurzfassung: To solve multi-objective problems, multiple reward signals are often scalarized into a single value and further processed using established single-objective problem solving techniques. While the field of multi-objective optimization has made many advances in applying scalarization techniques to obtain good solution trade-offs, the utility of applying these techniques in the multi-objective multi-agent learning domain has not yet been thoroughly investigated. Agents learn the value of their decisions by linearly scalarizing their reward signals at the local level, while acceptable system wide behaviour results. However, the non-linear relationship between weighting parameters of the scalarization function and the learned policy makes the discovery of system wide trade-offs time consuming. Our first contribution is a thorough analysis of well known scalarization schemes within the multi-objective multi-agent reinforcement learning setup. The analysed approaches intelligently explore the weight-space in order to find a wider range of system trade-offs. In our second contribution, we propose a novel adaptive weight algorithm which interacts with the underlying local multi-objective solvers and allows for a better coverage of the Pareto front. Our third contribution is the experimental validation of our approach by learning biobjective policies in self-organising smart camera networks. We note that our algorithm (i) explores the objective space faster on many problem instances, (ii) obtained solutions that exhibit a larger hypervolume, while (iii) acquiring a greater spread in the objective space.
Schlagworte:
Publikationstyp: Beitrag in Sammelwerk (Autorenschaft)
Art der Veröffentlichung Online Publikation
Erschienen in: Proceedings of the IEEE International Joint Conference on Neural Networks (IJCNN´2014), July 6-11 2014, Beijing, China
Proceedings of the IEEE International Joint Conference on Neural Networks (IJCNN´2014), July 6-11 2014, Beijing, China
zur Publikation
 ( )
Erscheinungdatum: 2014
Titel der Serie: -
Bandnummer: -
Erstveröffentlichung: Ja
Version: -
Seite: S. 2306 - 2314

Identifikatoren

ISBN: -
ISSN: -
DOI: http://dx.doi.org/10.1109/IJCNN.2014.6889637
AC-Nummer: -
Homepage: https://pervasive.aau.at/publications/pdf/Vmoffaert_wcci2014.pdf
Open Access
  • Kein Open-Access

Zuordnung

Organisation Adresse
Fakultät für Technische Wissenschaften
 
Institut für Vernetzte und Eingebettete Systeme
Universitätsstraße 65-67
9020  Klagenfurt am Wörthersee
Österreich
  -993640
   kornelia.lienbacher@aau.at
http://nes.aau.at/
zur Organisation
Universitätsstraße 65-67
AT - 9020  Klagenfurt am Wörthersee

Kategorisierung

Sachgebiete
  • 1020 - Informatik
  • 2020 - Elektrotechnik, Elektronik, Informationstechnik
Forschungscluster Kein Forschungscluster ausgewählt
Peer Reviewed
  • Ja
Publikationsfokus
  • Science to Science (Qualitätsindikator: I)
Klassifikationsraster der zugeordneten Organisationseinheiten:
Arbeitsgruppen Keine Arbeitsgruppe ausgewählt

Kooperationen

Keine Kooperationspartner ausgewählt

Beiträge der Publikation

Keine verknüpften Publikationen vorhanden