def __init__(self, args):
"""
我们人为生成一些上下文来模拟
:param args: 臂个数参数，以及各个臂的穿越参数等
"""
Policy.__init__(self, args)
self.alpha = args[1]
self.travel_args = args[2:] # 穿越过来的臂均值参数
self.d = 3 # 上下文维度
self.A = np.array([np.identity(self.d) for _ in range(self.n_bandits)])
self.b = np.array([np.zeros(self.d) for _ in range(self.n_bandits)])
self.context = None
def __init__(self, args):
Policy.__init__(self, args)
self.try_perSlot = int(args[1])
def __init__(self, args):
Policy.__init__(self, args)
self.temperature = args[1] # 降火参数，温度越高，分子越随机，成为气体；低温的时候有序排列，成为固体
self.anneal = args[2] > 0.0
def __init__(self, args):
Policy.__init__(self, args)
def __init__(self, args):
Policy.__init__(self, args)
self.anneal = args[2] > 0.0
self.decay = args[1]
def __init__(self, args):
Policy.__init__(self, args)
# 记录每个臂的beta分布参数
self.betaArgs = [[args[1], args[2]] for _ in range(self.n_bandits)]
def __init__(self, args):
Policy.__init__(self, args)
self.gamma = args[1]
self._weights = np.array([1] * self.n_bandits)
self._probs = None
def __init__(self, args):
Policy.__init__(self, args)
self.squared_reward = [0.0] * self.n_bandits