def hmin_alpha_beta(state, player, pits, initial_max, alpha, beta, d):
global count
count = count + 1
if d == 0: # Reached depth threshold, evaluate heuristic
return evaluate(state, pits, initial_max)
if terminal_test(state, pits):
return utility(state, pits, initial_max)
u = math.inf # max possible utility
actions = available_actions(state, player, pits)
for a in actions: # Check each possible action for minimmum utility
player_copy = player
state_copy = [row[:] for row in state]
result_player, result_state = action(player_copy, a, state_copy, pits)
if result_player == player: # Run min_value if next player is MIN
u = min([
u,
hmin_alpha_beta(result_state, result_player, pits, initial_max,
alpha, beta, d - 1)
])
if u <= alpha:
return u
beta = min(beta, u)
else: # Run max_value if next player is MAX
u = min([
u,
hmax_alpha_beta(result_state, result_player, pits, initial_max,
alpha, beta, d - 1)
])
return u
def hmaxvalue(state, player, pits, p_o, d):
global count
count = count + 1
v = -(math.inf)
player_initial = player
if d == 0: # Reached depth threshold, evaluate heuristic
return evaluate(state, pits, p_o)
if not terminal_test(state, pits):
actions = available_actions(state, player, pits)
for move in actions:
new_state = [row[:] for row in state]
player, new_state_1 = action(
player_initial, move, new_state,
pits) # Check resulting state of each action
if player == player_initial: # If same player, perform hmaxvalue again
v = max(v, hmaxvalue(new_state_1, player, pits, p_o, d - 1))
else: # If player changes, perform hminvalue
v = max(v, hminvalue(new_state_1, player, pits, p_o, d - 1))
else: # Terminal case reached, return utility
return utility(state, pits, p_o)
return v
def max_alpha_beta(state, player, pits, initial_max, alpha, beta):
global count
count = count + 1
if terminal_test(state, pits):
return utility(state, pits, initial_max)
u = -math.inf # min possible utility
actions = available_actions(state, player, pits)
for a in actions: # Check each possible action for maximum utility
player_copy = player
state_copy = [row[:] for row in state]
result_player, result_state = action(player_copy, a, state_copy, pits)
if result_player == player: # Run max_value if next player is MAX
u = max([
u,
max_alpha_beta(result_state, result_player, pits, initial_max,
alpha, beta)
])
if u >= beta:
return u
alpha = max(alpha, u)
else: # Run min_value if next player is MIN
u = max([
u,
min_alpha_beta(result_state, result_player, pits, initial_max,
alpha, beta)
])
return u
def min_value(state, player, pits, initial_max):
global count
count = count + 1
if terminal_test(state, pits):
return utility(state, pits, initial_max)
u = math.inf # Set maximum possible utility
actions = available_actions(state, player, pits)
for a in actions: # Check each possible action for minimmum utility
player_copy = player
state_copy = [row[:] for row in state]
result_player, result_state = action(player_copy, a, state_copy, pits)
if result_player == player: # Run min_value if next player is MIN
u = min(
[u,
min_value(result_state, result_player, pits, initial_max)])
else: # Run max_value if next player is MAX
u = min(
[u,
max_value(result_state, result_player, pits, initial_max)])
return u
