本文整理汇总了Python中pybrain.rl.agents.LearningAgent.newEpisode方法的典型用法代码示例。如果您正苦于以下问题：Python LearningAgent.newEpisode方法的具体用法？Python LearningAgent.newEpisode怎么用？Python LearningAgent.newEpisode使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类pybrain.rl.agents.LearningAgent的用法示例。
在下文中一共展示了LearningAgent.newEpisode方法的1个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: Team
# 需要导入模块: from pybrain.rl.agents import LearningAgent [as 别名]
# 或者: from pybrain.rl.agents.LearningAgent import newEpisode [as 别名]
class Team(object):
def __init__(self, living, task, learner = ENAC()):
self.living = living
self.task = task
self.last_reward = 0
self.agent = LearningAgent(self.living.brain, learner)
self.oldparams = self.living.brain.params
def Interaction(self):
self.agent.integrateObservation(self.task.getObservation())
self.task.performAction(self.agent.getAction())
self.last_reward = self.task.getReward()
self.agent.giveReward(self.last_reward)
finished = self.task.isFinished()
if finished:
#print task.cumreward
self.agent.newEpisode()
self.task.reset()
return self.last_reward, finished
def Learn(self, episodes = 1):
self.agent.learn(episodes)
self.agent.reset()
newparams = self.living.brain.params.copy() #get_all_weights(eater.brain)[:]
dif = 0
j = 0
for i in newparams:
dif += (self.oldparams[j] - newparams[j])**2
j += 1
self.oldparams = newparams
return dif