`project_dis_restoration.NN_model`¶

Module Contents¶

Classes¶

`LossHistory`
`CNN_TieLine`
`DNN_TieLine`
`DNN_VarCon`

Attributes¶

`ALPHA`
`GAMMA`
`LEARNING_RATE`
`TRAIN_HIST_SIZE`
`EXPERIENCE_MEMORY`
`BATCH_SIZE`
`MEMORY_SIZE`
`EXPLORATION_MAX`
`EXPLORATION_MIN`
`EXPLORATION_DECAY`

project_dis_restoration.NN_model.ALPHA = 0.1¶

project_dis_restoration.NN_model.GAMMA = 0.95¶

project_dis_restoration.NN_model.LEARNING_RATE = 0.001¶

project_dis_restoration.NN_model.TRAIN_HIST_SIZE = 100000¶

project_dis_restoration.NN_model.EXPERIENCE_MEMORY = 100000¶

project_dis_restoration.NN_model.BATCH_SIZE = 32¶

project_dis_restoration.NN_model.MEMORY_SIZE = 10000000¶

project_dis_restoration.NN_model.EXPLORATION_MAX = 1.0¶

project_dis_restoration.NN_model.EXPLORATION_MIN = 0.01¶

project_dis_restoration.NN_model.EXPLORATION_DECAY = 0.999¶

class project_dis_restoration.NN_model.LossHistory¶

Bases: keras.callbacks.Callback

on_train_begin(self, logs={})¶

on_batch_end(self, batch, logs={})¶

class project_dis_restoration.NN_model.CNN_TieLine(observation_space, action_space)¶

collect(self, s, expert_a)¶

end_collect(self)¶

train(self)¶

predict(self, x, batch_size=1)¶: predict on (a batch of) x

class project_dis_restoration.NN_model.DNN_TieLine(observation_space, action_space)¶

collect(self, s, expert_a, other=None)¶

end_collect(self)¶

train(self)¶

predict(self, x, batch_size=1)¶: predict on (a batch of) x

remember(self, state, action, reward, next_state, done)¶

experience_replay(self)¶

class project_dis_restoration.NN_model.DNN_VarCon(observation_space, action_space)¶

collect(self, s, expert_a)¶

end_collect(self)¶

train(self)¶

predict(self, x, batch_size=1)¶: predict on (a batch of) x