Afzal Upal, Rik Warren, Ganesh Alakke
This paper advocates a new methodology for studying computational cultural dynamics. It argues that knowledge-rich agent-based social simulations are needed to understand and model the cultural dynamics of natural and artificial societies. The methodology is illu strated with the help of Multiagent Wumpus World (MWW) testbed. We also presents results of experiments conducted using a version of MWW.
Subjects: 7.1 Multi-Agent Systems; 12.1 Reinforcement Learning
Submitted: Jun 20, 2008