Abstract | ||
---|---|---|
In multi-agent systems, the presence of learning agents can cause the environment to be non-Markovian from an agent's perspective thus violating the property that traditional single-agent learning methods rely upon. This paper formalizes some known intuition about concurrently learning agents by providing formal conditions that make the environment non-Markovian from an independent (non-communicative) learner's perspective. New concepts are introduced like the divergent learning paths and the observability of the effects of others' actions. To illustrate the formal concepts, a case study is also presented. These findings are significant because they both help to understand failures and successes of existing learning algorithms as well as being suggestive for future work. |
Year | DOI | Venue |
---|---|---|
2011 | 10.3233/KES-2010-0206 | KES Journal |
Keywords | Field | DocType |
future work,divergent learning path,multi-agent system,environment non-markovian,formal condition,formal concept,traditional single-agent,known intuition,new concept,case study,independent learner,reinforcement learning,machine learning,multi agent system | Robot learning,Algorithmic learning theory,Instance-based learning,Active learning,Active learning (machine learning),Computer science,Synchronous learning,Artificial intelligence,Sequence learning,Machine learning,Proactive learning | Journal |
Volume | Issue | ISSN |
15 | 1 | 1327-2314 |
Citations | PageRank | References |
15 | 1.11 | 26 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Guillaume J. Laurent | 1 | 97 | 12.60 |
Laëtitia Matignon | 2 | 88 | 9.43 |
N. Le Fort-Piat | 3 | 46 | 4.25 |