Eduardo Rodrigues Gomes Ryszard Kowalczyk Modelling the dynamics of multiagent Q-learning with ε-greedy exploration. 1181-1182 2009 AAMAS (2) http://doi.acm.org/10.1145/1558109.1558202 conf/atal/2009-2 db/conf/atal/aamas2009-2.html#GomesK09