def take_action_pre_train(self, phase_time_now):
current_phase_number = self.get_current_phase()
rewards_detail_dict_list = []
if (self.current_phase_duration <
phase_time_now[current_phase_number]):
action = 0
else:
action = 1
for i in range(self.para_set.MIN_ACTION_TIME):
action_in_second = 0
current_phase_number = self.get_current_phase()
if action == 1 and i == 0:
action_in_second = 1
self.current_phase, self.current_phase_duration, self.vehicle_dict = map_computor.run(
action=action_in_second,
current_phase=current_phase_number,
current_phase_duration=self.current_phase_duration,
vehicle_dict=self.dic_vehicles,
rewards_info_dict=self.para_set.REWARDS_INFO_DICT,
f_log_rewards=self.f_log_rewards,
rewards_detail_dict_list=rewards_detail_dict_list
) # run 1s SUMO
reward = self.cal_reward_from_list(rewards_detail_dict_list)
#self.update_vehicles()
self.update_state()
return reward, action
def take_action(self, action):
current_phase_number = self.get_current_phase()
rewards_detail_dict_list = []
if (self.current_phase_duration < self.para_set.MIN_PHASE_TIME[current_phase_number]):
action = 0
for i in range(self.para_set.MIN_ACTION_TIME):
action_in_second = 0
current_phase_number = self.get_current_phase()
if action == 1 and i == 0:
action_in_second = 1
# print("test", rewards_detail_dict_list)
self.current_phase, self.current_phase_duration, self.vehicle_dict = map_computor.run(action=action_in_second,
current_phase=current_phase_number,
current_phase_duration=self.current_phase_duration,
vehicle_dict=self.dic_vehicles,
rewards_info_dict=self.para_set.REWARDS_INFO_DICT,
f_log_rewards=self.f_log_rewards,
rewards_detail_dict_list=rewards_detail_dict_list) # run 1s SUMO
#reward, reward_detail_dict = self.cal_reward(action)
# print("test1", rewards_detail_dict_list)
reward = self.cal_reward_from_list(rewards_detail_dict_list)
#self.update_vehicles()
self.update_state()
return reward, action
def take_action(self, action, phase):
current_phase_number = self.get_current_phase()
rewards_detail_dict_list = []
if int(action) == int(phase):
changed = False
else:
changed = True
# 현재 신호가 최소 지속 시간을 넘지 않았다면 a
# MIN_ACTION_TIME 까지 돌아라(5까지)
for i in range(self.ParaSet.MIN_ACTION_TIME):
# action time 동안에
action_in_second = self.get_current_phase()
current_phase_number = self.get_current_phase()
#### 이 부분 바꿔야함!###
# a가 바꾸는 거라면 일단 첫 액션은 바꾸는 걸로 해라
if changed and i == 0:
action_in_second = action
# 현재 상황에서 actioninsecond을 주었을 때의 결과
self.current_phase, self.current_phase_duration, self.vehicle_dict = map_computor.run(
action=action_in_second,
current_phase=current_phase_number,
current_phase_duration=self.current_phase_duration,
vehicle_dict=self.dic_vehicles,
rewards_info_dict=self.ParaSet.REWARDS_INFO_DICT,
f_log_rewards=self.f_log_rewards,
changed=changed,
rewards_detail_dict_list=rewards_detail_dict_list
) # run 1s SUMO
changed = False
reward = self.cal_reward_from_list(rewards_detail_dict_list)
self.update_state()
return reward, action
