Technical Note: \cal Q-Learning
Machine Learning
Rational and convergent learning in stochastic games
IJCAI'01 Proceedings of the 17th international joint conference on Artificial intelligence - Volume 2
| Hi-index | 0.00 |
Norms are a valuable mechanism for establishing coherent cooperative behaviour in decentralised systems in which no central authority exists. In this context, Axelrod'sseminal model of norm establishment in populations of self interested individuals [1] is important in providing insight into the mechanisms needed to support this. However, Axelrod'smodel suffers from significant limitations: it adopts an evolutionary approach, and assumes that information is available to all agents in the system. In particular, the model assumes that the private strategies of individuals are available to others, and that agents are omniscient in being aware of all norm violations and punishments. Because this is an unreasonable expectation, the approach does not lend itself to modelling real world systems such as peer-to-peer networks. In response, this paper proposes alternatives to Axelrod's model, by replacing the evolutionary approach, enabling agents to learn, and by restricting the met punishment of agents to only those where the original defection is perceived, in order to be able to apply the model to real-world domains.