def test():
"""
Testing function : Training for a given game and 2 agents, and Battle between both agents
"""
#############################################
################ Choose Game ################
#############################################
#game = Soccer()
#game = RockPaperScissors()
game = GridWorld()
playerA, playerB = game.players()[0], game.players()[1] #player ID
#############################################
### Choose Player 1 and training opponent ###
#############################################
player1 = WoLF_PHC_Agent(game, playerA)
opponent1 = WoLF_PHC_Agent(game, playerB)
#############################################
### Choose Player 2 and training opponent ###
#############################################
player2 = Minimax_Q_Agent(game, playerB)
opponent2 = Random_Agent(game, playerA)
#############################################
############## Train Policies ###############
#############################################
nb_iterations = 500000
timestamp = 1000
start_time = time.time()
policy1, policyb = Scheduler(game, nb_iterations, timestamp, player1,
opponent1)
print("Learning Time Player 1: ", time.time() - start_time)
start_time = time.time()
policyc, policy2 = Scheduler(game, nb_iterations, timestamp, opponent2,
player2)
print("Learning Time Player 2: ", time.time() - start_time)
#policy 1 : distances between Nash 1 and 2
optimal_Nash1_player0 = GridWorld_Nash1_Player0_Agent(game)
optimal_Nash1_player0.compute_policy()
optimal_Nash2_player0 = GridWorld_Nash2_Player0_Agent(game)
optimal_Nash2_player0.compute_policy()
d10 = []
d20 = []
for i in range(len(policy1)):
d10.append(distance(policy1[i], optimal_Nash1_player0.pi))
d20.append(distance(policy1[i], optimal_Nash2_player0.pi))
#policy 2 : distances between Nash 1 and 2
optimal_Nash1_player1 = GridWorld_Nash1_Player1_Agent(game)
optimal_Nash1_player1.compute_policy()
optimal_Nash2_player1 = GridWorld_Nash2_Player1_Agent(game)
optimal_Nash2_player1.compute_policy()
d11 = []
d21 = []
for i in range(len(policy2)):
d11.append(distance(policy2[i], optimal_Nash1_player1.pi))
d21.append(distance(policy2[i], optimal_Nash2_player1.pi))
#plot :
plt.plot(d10, 'b')
plt.plot(d20, 'b--')
plt.plot(d11, 'r')
plt.plot(d21, 'r--')
plt.show()
#Battle :
print("Battle")
nbplay = 1000
affrontement(game, policy1[-1], policy2[-1], nbplay)
return (policy1, policy2)
#test()
