def __init__(self, name, num_states, num_actions, epsilon=0.3, gamma=0.99, alpha=0.95):
self.controller = ActionValueTable(num_states, num_actions)
self.controller.initialize(np.random.rand(num_states * num_actions))
self.learner = Q(gamma=gamma, alpha=alpha)
self.learner.batchMode = False
self.learner.explorer.epsilon = epsilon
LearningAgent.__init__(self, self.controller, self.learner)
Agent.__init__(self, name)
def __init__(self, outdim, n_actions, random_state, rl_params):
""" RL agent
"""
module = SparseActionValueTable(n_actions, random_state)
module.initialize(0.0)
learner = EpisodeQ(alpha=rl_params.q_alpha,
w=rl_params.q_w,
gamma=rl_params.q_gamma)
learner.explorer = EGreedyExplorer(random_state,
epsilon=rl_params.exp_epsilon,
decay=rl_params.exp_decay)
LearningAgent.__init__(self, module, learner)
def __init__(self, _id, module, learner=None): #define variaveis da class self.id = _id self.horizontal_edge = lane.getEdgeID(trafficlights.getControlledLanes(self.id)[0]) self.vertical_edge = lane.getEdgeID(trafficlights.getControlledLanes(str(_id))[2]) #define variaveis da classe pai self.horizontalLoad = [] self.verticalLoad = [] self.averageHorizontal = [] self.averageVertical = [] self.nextAction = None self.expectedReward = None self.tolerance = None LearningAgent.__init__(self, module, learner)
def __init__(self,
name,
num_states,
num_actions,
epsilon=0.3,
gamma=0.99,
alpha=0.95):
self.controller = ActionValueTable(num_states, num_actions)
self.controller.initialize(np.random.rand(num_states * num_actions))
self.learner = Q(gamma=gamma, alpha=alpha)
self.learner.batchMode = False
self.learner.explorer.epsilon = epsilon
LearningAgent.__init__(self, self.controller, self.learner)
Agent.__init__(self, name)
def __init__(self, module, learner = None):
'''
Constructor
'''
LearningAgent.__init__(self, module, learner)
self.__rules=[]
self.__states={}
self.__input={}
self.__buffer={}
# self.__rules.append(BackOffRule())
self.__rules.append(BackOffRule2())
self.__rules.append(LocomotionPrimitives())
self.__states["driveBackStartTime"]=AgentMind.__driveBackStartTime
self.__states["__lostTrackTurnStartTime"]=AgentMind.__lostTrackTurnStartTime
def __init__(self, module, learner=None):
'''
Constructor
'''
LearningAgent.__init__(self, module, learner)
self.__rules = []
self.__states = {}
self.__input = {}
self.__buffer = {}
# self.__rules.append(BackOffRule())
self.__rules.append(BackOffRule2())
self.__rules.append(LocomotionPrimitives())
self.__states["driveBackStartTime"] = AgentMind.__driveBackStartTime
self.__states[
"__lostTrackTurnStartTime"] = AgentMind.__lostTrackTurnStartTime
def __init__(self, x, y, brain, learner, env):
LearningAgent.__init__(self, brain.net, learner)
self.cellType = 3
self.brain = brain
self.module = brain.net
self.learner = learner
self.env = env
self.color = cell.BLACK
self.x = x
self.y = y
self.num_interactions = 0
self.age = 0
self.colddown = 0
self.speed = self.Speeds[0]
self.energy = self.MaxEnergy
self.food_sensor = 0
self.hunger_sensor = 0
self.target = [-1, -1]
def __init__(self, x, y, brain, learner, env):
LearningAgent.__init__(self, brain.net, learner)
self.cellType = 3
self.brain = brain
self.module = brain.net
self.learner = learner
self.env = env
self.color = cell.BLACK
self.x = x
self.y = y
self.num_interactions = 0
self.age = 0
self.colddown = 0
self.speed = self.Speeds[0]
self.energy = self.MaxEnergy
self.food_sensor = 0;
self.hunger_sensor = 0;
self.target = [-1, -1]