def play_versus_learner(strategy, value_map, verbose=True, explore=0.1):
state0 = (0, 0, 0, 0, 0, 0, 0, 0, 0)
alpha = 0.5
while True:
state1 = strategy(1, state0)
if is_winner(1, state1):
value_map[state0] = 0
if verbose: print("\033[1;31;40m player 1 wins \n")
return -1
if is_finished(state1):
value_map[state0] = 0.5
if verbose: print("\033[1;33;40m draw \n")
return 0
state2 = player_move_learner(2, state1, value_map, False, explore)
#update after every player 2 move=>
value_map[state0] = (
value_map.get(state0, 0.5) + alpha *
(value_map.get(state2, 0.5) - value_map.get(state0, explore)))
if is_winner(2, state2):
value_map[state2] = 1
if verbose: print("\033[1;34;40m player 2 wins \n")
return 1
if is_finished(state2):
value_map[state2] = 0.5
if verbose: print("\033[1;33;40m draw \n")
return 0
# reset to iterate again
state0 = state2
def random_versus_never_lose(verbose=True):
state0 = (0, 0, 0, 0, 0, 0, 0, 0, 0)
while True:
state1 = player_move_random(1, state0)
if is_winner(1, state1):
if verbose: print("\033[1;31;40m player 1 wins \n")
return -1
if is_finished(state1):
if verbose: print("\033[1;33;40m draw \n")
return 0
state2 = player_move_never_lose(2, state1)
if is_winner(2, state2):
if verbose: print("\033[1;34;40m player 2 wins \n")
return 1
if is_finished(state2):
if verbose: print("\033[1;33;40m draw \n")
return 0
# reset to iterate again
state0 = state2
def next_play(player, state, strategy, verbose=False):
next_state = strategy(player, state)
if verbose: print(state_str(next_state) + "\n-----")
if is_winner(player, next_state):
if verbose: print("player %i wins" % player)
return (player, None)
elif is_finished(next_state):
if verbose: print("draw")
return (0, None)
else:
return (None, next_state)
def two_learners(value_map_1, value_map_2, explore1, explore2, verbose=False):
state0 = (0, 0, 0, 0, 0, 0, 0, 0, 0)
state1 = player_move_learner(1, state0, value_map_1, False, explore1)
alpha = 0.5 # learning rate is set to 0.5
while True:
state2 = player_move_learner(2, state1, value_map_2, False, explore2)
#update after every player 2 move
value_map_2[state0] = (
value_map_2.get(state0, 0.5) + alpha *
(value_map_2.get(state2, 0.5) - value_map_2.get(state0, explore2)))
if is_winner(2, state2):
value_map_2[state2] = 1
if verbose: print("\033[1;34;40m player 2 wins \n")
return -1
if is_finished(state2):
value_map_2[state2] = 0.5
if verbose: print("\033[1;33;40m draw \n")
return 0
state3 = player_move_learner(1, state2, value_map_1, False, explore1)
value_map_1[state1] = (
value_map_1.get(state1, 0.5) + alpha *
(value_map_1.get(state3, 0.5) - value_map_1.get(state1, explore1)))
if is_winner(1, state3):
value_map_1[state0] = 0
if verbose: print("\033[1;31;40m player 1 wins \n")
return 1
if is_finished(state3):
value_map_1[state0] = 0.5
if verbose: print("\033[1;33;40m draw \n")
return 0
# reset to iterate again
state0 = state2
state1 = state3