Research Article

Limiting Dynamics for Q-Learning with Memory One in Symmetric Two-Player, Two-Action Games

Figure 2

(a) Phase diagram for the possible BRNs in the prisoner’s dilemma, (b) phase diagram for the possible equilibria in the prisoner’s dilemma (only node 0 in region 1, nodes 0 and 1 in region 2, and nodes 0, 1, and 9 in region 3). For both plots, we have , and .
(a)
(b)