Home |  ENGLISH |  Kontakt |  Impressum |  Anmelden |  KIT

Inproceedings1709

Aus Aifbportal

Wechseln zu: Navigation, Suche

(This page contains COinS metadata)

Learning to Control the Emergent Behaviour of a Multi-Agent System


Urban Richter, Moez Mnif



Published: 2008 Mai
Herausgeber: Franziska Klügl, Karl Tuyls, and Sandip Sen
Buchtitel: Proceedings of the 2008 Workshop on Adaptive Learning Agents and Multi-Agent Systems at AAMAS 2008 (ALAMAS+ALAg 2008)
Seiten: 33-40
Referierte Veröffentlichung
BibTeX

Kurzfassung
Organic Computing (OC) has the vision of addressing the challenges of complex distributed systems by making them more life-like (organic), i. e., endowing them with abilities such as self-organisation, self-configuration, self-repair, or adaptation. This can only be achieved by giving the system elements adequate degrees of freedom, which may result in an emergent behaviour, which can be positive (desired) as well as negative (undesired). In this context, the so-called observer/controller architecture has become widespread in the OC community, as a design paradigm to assure the fulfilment of system goals (given by the developer or user). An observer/controller loop enables adequate reactions to control the – sometimes completely unexpected – undesired emerging global behaviour resulting from local agents' behaviour. Therefore, organic systems need the ability to quantify the system status as well as the ability to learn (because in most cases the developer cannot predict all system states).

In this paper, a nature-inspired multi-agent scenario is taken to validate a generic observer/controller architecture, which has been designed as part of an intended organic framework – providing generic toolbox mechanisms to observe, analyse, and control emergent behaviour in self-organising systems. In particular, we discuss the learning ability of the 2-level learning architecture of the controller by presenting first experimental results of different learning methods and strategies.


Projekt

OCCSQE


Forschungsgruppe

Effiziente Algorithmen


Forschungsgebiet
Agenten Systeme, Maschinelles Lernen, Organic Computing