mia/nim
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Implementiere best_future_reward

Browse source
This commit is contained in:
mia 2026-04-23 10:00:43 +02:00
parent 933d0c5d27
commit e4fca60692
3 changed files with 11 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

BIN
__pycache__/nim.cpython-313.pyc
View file

Binary file not shown.

16
nim.py
View file

@ -95,15 +95,17 @@ class NimAI():
float: The highest Q-value among available actions.
Returns 0 if no actions are available.
"""
state = tuple(state)
# actions = []
# for q in self.q.key:
# if q[0] == state:
# for q in self.q:
# print(q)
# if q[0] == tuple(state):
# actions.append(q[1])
actions = tuple([key[1] for key in self.q.keys() if key[0] == state])
try:
return max([q for q in self.q[tuple(state), actions]])
except:
return 0
actions = [key[1] for key in self.q if key[0] == state]
# print(actions)
# print(self.q[state, action] for action in actions)
# print(max(self.q[state, action] for action in actions))
return 0 if actions == [] else max(self.q[state, action] for action in actions)
def choose_action(self, state, epsilon=True):
"""

2
test.py
View file

@ -19,6 +19,8 @@ def test_update_q_value(ai):
def test_best_future_reward(ai):
print("\n--- Testing best_future_reward ---")
print(ai.best_future_reward([1,1,1,0]))
print(ai.best_future_reward([1,1,1,1]))
def test_choose_action(ai):