doc(FrozenLake_tuto): update policy exploitation logic to handle vari…

…able sets of maximum Q-values (#1037)

docs/tutorials/training_agents/FrozenLake_tuto.py

@@ -161,12 +161,10 @@ def choose_action(self, action_space, state, qtable):
         # Exploitation (taking the biggest Q-value for this state)
         else:
             # Break ties randomly
-            # If all actions are the same for this state we choose a random one
-            # (otherwise `np.argmax()` would always take the first one)
-            if np.all(qtable[state, :]) == qtable[state, 0]:
-                action = action_space.sample()
-            else:
-                action = np.argmax(qtable[state, :])
+            # Find the indices where the Q-value equals the maximum value
+            # Choose a random action from the indices where the Q-value is maximum
+            max_ids = np.where(qtable[state, :] == max(qtable[state, :]))[0]
+            action = rng.choice(max_ids)
         return action