def test_reset(self, pev, CS_list):
self.path_info = []
self.CS_list = CS_list
self.pev = pev
self.sim_time = self.pev.t_start
self.path_info = rt.sim_main_first_time_check(self.pev, self.CS_list,
self.graph,
self.action_size)
state = [self.pev.source, self.pev.curr_SOC]
for path in self.path_info:
cs, pev_SOC, front_path, total_d_time, waiting_time, charging_time = path
state += [total_d_time, waiting_time, charging_time]
state = np.reshape(state, [1, self.state_size])
return state, source, destination
def reset(self):
t_start = np.random.uniform(0, 1200)
soc = np.random.uniform(0.3, 0.5)
while soc <= 0.0 or soc > 1.0:
soc = np.random.uniform(0.3, 0.5)
self.graph.source_node_set = list(self.graph.source_node_set)
self.graph.destination_node_set = list(self.graph.destination_node_set)
source = self.graph.source_node_set[np.random.random_integers(
0,
len(self.graph.source_node_set) - 1)]
while source in self.graph.cs_info.keys():
source = self.graph.source_node_set[np.random.random_integers(
0,
len(self.graph.source_node_set) - 1)]
destination = self.graph.source_node_set[np.random.random_integers(
0,
len(self.graph.source_node_set) - 1)]
while destination in self.graph.cs_info.keys():
destination = self.graph.source_node_set[np.random.random_integers(
0,
len(self.graph.source_node_set) - 1)]
self.path_info = []
self.pev = EV(e, t_start, soc, source, destination)
self.sim_time = self.pev.t_start
self.path_info = rt.sim_main_first_time_check(self.pev, self.CS_list,
self.graph,
self.action_size)
# print('\npev soc', self.pev.SOC)
state = [self.pev.source, self.pev.SOC]
for path in self.path_info:
cs, pev_SOC, front_path, total_d_time, waiting_time, charging_time = path
state += [total_d_time, waiting_time, charging_time]
state = np.reshape(state, [1, self.state_size])
return state, source, destination
def step(self, action, done):
cs, pev_SOC, front_path, total_d_time, waiting_time, charging_time = self.path_info[
action]
# print(self.sim_time, self.pev.curr_location, cs.id, pev_SOC)
self.target = cs
if pev_SOC <= 0.0:
print('error soc', pev_SOC)
done = 1
reward = -5
return np.zeros((1, self.state_size)), -1, reward, done
if len(front_path) > 1:
next_node = front_path[1]
self.sim_time, time = rt.update_ev(self.pev, self.graph,
self.pev.curr_location,
next_node, self.sim_time)
if self.sim_time == 0 and time == 0:
print('time idx error')
done = 1
reward = -5
return np.zeros((1, self.state_size)), -1, reward, done
done = 0
reward = -1 * (time)
self.pev.curr_location = next_node
self.path_info = rt.sim_main_first_time_check(
self.pev, self.CS_list, self.graph, self.action_size)
next_state = [self.pev.curr_location, self.pev.curr_SOC]
for path in self.path_info:
cs, pev_SOC, front_path, total_d_time, waiting_time, charging_time = path
next_state += [total_d_time, waiting_time, charging_time]
next_state = np.reshape(next_state, [1, self.state_size])
return next_state, next_node, reward, done
elif self.pev.curr_location == cs.id:
self.pev.before_charging_SOC = self.pev.curr_SOC
self.pev.cscharingenergy = self.pev.maxBCAPA * self.pev.req_SOC - self.pev.curr_SOC * self.pev.maxBCAPA
self.pev.cschargingcost = self.pev.cscharingenergy * cs.price[int(
self.sim_time / 5)]
self.pev.curr_SOC = self.pev.req_SOC
self.pev.cschargingtime = charging_time
self.pev.cschargingwaitingtime = waiting_time
self.pev.charged = 1
self.pev.cs = cs
self.sim_time += charging_time * 60
self.sim_time += waiting_time * 60
# print(waiting_time, waiting_time * 60)
self.pev.csdrivingtime = self.pev.totaldrivingtime
self.pev.csdistance = self.pev.totaldrivingdistance
self.pev.cschargingwaitingtime = self.pev.cschargingwaitingtime
self.pev.cschargingtime = self.pev.cschargingtime
self.pev.cssoc = self.pev.curr_SOC
done = 1
reward = -1 * (waiting_time + charging_time)
print(done, '충전소야')
return np.zeros((1, self.state_size)), -1, reward, done
else:
print("???")
input()