def at_least_n_times_exploration(actions, utilities, temperature, action_counter):
utilities = [utilities[x] for x in actions]
for i, utility in enumerate(utilities):
if action_counter[actions[i]] < min_n:
utilities[i] = optimistic_reward
d = dict(zip(actions, utilities))
uf = lambda action: d[action]
return argmax(actions, uf)
def at_least_n_times_exploration(actions, utilities, temperature,
action_counter):
utilities = [utilities[x] for x in actions]
for i, utility in enumerate(utilities):
if action_counter[actions[i]] < min_n:
utilities[i] = optimistic_reward
d = dict(zip(actions, utilities))
uf = lambda action: d[action]
return argmax(actions, uf)
def test_random_tie(self):
a = 0
for x in range(100):
if argmax(['a', 'b', 'c'], lambda x: self.d[x]) == 'a':
a += 1
self.assertTrue(25 < a < 75)
def test_return_max(self):
self.assertEqual('a', argmax(['a', 'b'], lambda x: self.d[x]))
