- simplified debug flags
- added params class - cleaned up
This commit is contained in:
+16
-9
@@ -5,15 +5,22 @@ from numpy.random import uniform
|
||||
|
||||
|
||||
class MachinePlayer(APlayer):
|
||||
def __init__(self, mark='X', with_debug=False, values=None):
|
||||
class Params:
|
||||
def __init__(self, p_exp, alpha):
|
||||
self.p_exp = p_exp
|
||||
self.alpha = alpha
|
||||
|
||||
def __init__(self, mark, params: Params):
|
||||
APlayer.__init__(self, mark)
|
||||
self.p_exp = 0.2
|
||||
self.alpha = 0.1
|
||||
self.params = params
|
||||
self.values = {}
|
||||
self.with_debug = False
|
||||
|
||||
def set_debug(self, with_debug):
|
||||
self.with_debug = with_debug
|
||||
if values is None:
|
||||
self.init_values()
|
||||
else:
|
||||
self.values = values
|
||||
|
||||
def assign_values(self, values):
|
||||
self.values = values
|
||||
|
||||
def init_values(self):
|
||||
self.values = {}
|
||||
@@ -78,7 +85,7 @@ class MachinePlayer(APlayer):
|
||||
|
||||
next_move = best_move
|
||||
is_exp = False
|
||||
if uniform() < self.p_exp:
|
||||
if uniform() < self.params.p_exp:
|
||||
index = np.random.randint(len(moves))
|
||||
next_move = moves[index]
|
||||
is_exp = best_move != next_move
|
||||
@@ -101,7 +108,7 @@ class MachinePlayer(APlayer):
|
||||
if not is_exp and self.state_last is not None:
|
||||
v0 = self.get_value(self.state_last)
|
||||
v1 = self.get_value(self.state)
|
||||
d = v0 + self.alpha*(v1-v0)
|
||||
d = v0 + self.params.alpha*(v1-v0)
|
||||
if d > 0:
|
||||
self.set_value(self.state_last, d)
|
||||
if self.with_debug:
|
||||
|
||||
Reference in New Issue
Block a user