Experience Weighted Learning in Multiagent Systems

Algorithm 1

-learning.

(1)	repeat
(2)	i = 0
(3)	Initialize Q (s, a)
(4)	repeat
(5)	Choose an action A using policy derived from Q (e.g., -greedy)
(6)	Choose an opponent randomly
(7)	Take action A and observe R,
(8)
(9)
(10)	until S is terminal
(11)	i = i + 1.
(12)	until i = the total number of all the agents