def test_reset(self, pev, graph, CS_list):
self.graph = graph
self.path_info = []
self.CS_list = CS_list
self.pev = pev
self.sim_time = self.pev.t_start
self.timeIDX = int(self.sim_time / 5)
self.path_info = ta.get_feature_state(self.sim_time, self.pev,
self.CS_list, self.graph, NCS)
self.pev.path.append(self.pev.curr_location)
state = [
self.pev.curr_location / self.graph.num_node,
self.pev.destination / self.graph.num_node, self.pev.curr_SOC,
self.sim_time / 288
]
for path in self.path_info:
cs, req_soc, driving_cost, fpath, rpath, fp_dist, rp_dist, fp_dt, rp_dt, fp_w, r_w, wtime, ctime, cs_char_cost, home_char_cost = path
state += [
wtime, ctime, driving_cost / 10, cs_char_cost / 10,
home_char_cost / 10
]
state = np.reshape(state, [1, self.state_size])
return state, self.pev.source, self.pev.destination
def reset(self):
# self.graph.reset_traffic_info()
self.CS_list = reset_CS_info(self.graph)
t_start, soc = default_setting_random_value()
self.graph.source_node_set = list(self.graph.source_node_set)
source = self.graph.source_node_set[np.random.randint(
0,
len(self.graph.source_node_set) - 1)]
self.graph.destination_node_set = list(self.graph.destination_node_set)
destination = self.graph.destination_node_set[np.random.randint(
0,
len(self.graph.destination_node_set) - 1)]
while source == destination or destination in self.graph.cs_info:
destination = self.graph.destination_node_set[np.random.randint(
0,
len(self.graph.destination_node_set) - 1)]
self.path_info = []
self.pev = EV(e, t_start, soc, source, destination)
self.pev.path.append(self.pev.curr_location)
self.sim_time = self.pev.t_start
self.timeIDX = int(self.sim_time / 5)
self.path_info = ta.get_feature_state(self.sim_time, self.pev,
self.CS_list, self.graph, NCS)
# print('\npev soc', self.pev.SOC)
state = [
self.pev.curr_location / self.graph.num_node,
self.pev.destination / self.graph.num_node, self.pev.curr_SOC,
self.sim_time / 288
]
for path in self.path_info:
cs, req_soc, driving_cost, fpath, rpath, fp_dist, rp_dist, fp_dt, rp_dt, fp_w, r_w, wtime, ctime, cs_char_cost, home_char_cost = path
state += [
wtime, ctime, driving_cost / 10, cs_char_cost / 10,
home_char_cost / 10
]
state = np.reshape(state, [1, self.state_size])
return state, source, destination
def step(self, action):
cs, req_soc, driving_cost, fpath, rpath, fp_dist, rp_dist, fp_dt, rp_dt, fp_w, r_w, wtime, ctime, cs_char_cost, home_char_cost = self.path_info[
action]
self.pev.path.append(self.pev.curr_location)
if len(fpath) > 1:
next_node = fpath[1]
chargingprice = cs.price[int(self.sim_time / 5)]
distance_onetwo = self.graph.get_path_distance(fpath[0:2])
velocity_onetwo = self.graph.velocity(fpath[0], fpath[1],
int(self.sim_time / 5))
croad = self.pev.ECRate * distance_onetwo * chargingprice
troad = UNITtimecost * distance_onetwo / velocity_onetwo
road_cost = croad + troad
self.sim_time, time = ta.update_ev(self.pev, self.graph,
self.pev.curr_location,
next_node, self.sim_time)
if self.sim_time == 0 and time == 0:
done = 1
reward = -50
return np.zeros((1, self.state_size)), -1, reward, done
if self.pev.curr_SOC <= 0.0:
done = 1
reward = -50
return np.zeros((1, self.state_size)), -1, reward, done
done = 0
reward = -1 * road_cost
self.path_info = ta.get_feature_state(self.sim_time, self.pev,
self.CS_list, self.graph,
self.action_size)
next_state = [
self.pev.curr_location / self.graph.num_node,
self.pev.destination / self.graph.num_node, self.pev.curr_SOC,
self.sim_time / 288
]
for path in self.path_info:
cs, req_soc, driving_cost, fpath, rpath, fp_dist, rp_dist, fp_dt, rp_dt, fp_w, r_w, wtime, ctime, cs_char_cost, home_char_cost = path
next_state += [
wtime, ctime, driving_cost / 10, cs_char_cost / 10,
home_char_cost / 10
]
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.charged = 1
self.pev.cs = cs
self.pev.csid = cs.id
self.pev.req_SOC = req_soc
self.pev.rear_path = rpath
self.pev.cssoc = self.pev.curr_SOC
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.cschargingwaitingtime = cs.waittime[int(self.sim_time /
5)]
self.pev.cschargingstarttime = self.sim_time + self.pev.cschargingwaitingtime * 60
self.pev.cschargingprice = cs.price[int(
self.pev.cschargingstarttime / 5)]
self.pev.cschargingcost = self.pev.cscharingenergy * self.pev.cschargingprice
self.pev.csdrivingtime = self.pev.totaldrivingtime
self.pev.csdistance = self.pev.totaldrivingdistance
self.pev.cschargingwaitingtime = self.pev.cschargingwaitingtime
self.pev.cschargingtime = (
self.pev.cscharingenergy /
(cs.chargingpower * self.pev.charging_effi))
self.pev.curr_SOC = self.pev.req_SOC
self.sim_time += self.pev.cschargingwaitingtime * 60
self.sim_time += self.pev.cschargingtime * 60
self.pev.curr_time = self.sim_time
rpath = self.pev.rear_path
# print(rpath)
for i in range(len(rpath) - 1):
fnode = rpath[i]
tnode = rpath[i + 1]
# print(fnode, tnode, self.pev.curr_location)
self.pev.path.append(tnode)
self.sim_time, homedtime = ta.update_ev(
self.pev, self.graph, fnode, tnode, self.sim_time)
self.pev.homedrivingtime += homedtime
# print(self.pev.curr_location,self.pev.destination)
if self.pev.curr_location == self.pev.destination:
self.pev.homesoc = self.pev.curr_SOC
self.pev.homechargingstarttime = self.sim_time
self.pev.homechargingenergy = self.pev.maxBCAPA * self.pev.final_soc - self.pev.curr_SOC * self.pev.maxBCAPA
self.pev.homechargingtime = (
self.pev.homechargingenergy /
(cs.homechargingpower * self.pev.charging_effi))
self.pev.homedrivingdistance = self.graph.get_path_distance(
rpath)
self.pev.homechargingprice = cs.TOU_price[int(self.sim_time /
60)]
self.pev.homechargingcost = self.pev.homechargingenergy * self.pev.homechargingprice
self.pev.curr_SOC = self.pev.final_soc
self.sim_time += self.pev.homechargingtime
self.pev.curr_time = self.sim_time
self.pev.expense_time_part = (
self.pev.totaldrivingtime + self.pev.cschargingwaitingtime
+ self.pev.cschargingtime) * UNITtimecost
self.pev.expense_cost_part = self.pev.totaldrivingdistance * self.pev.ECRate * self.pev.cschargingprice + self.pev.cschargingcost + self.pev.homechargingcost
self.pev.totalcost = self.pev.expense_time_part + self.pev.expense_cost_part
done = 1
# reward = -1 * ((self.pev.cschargingwaitingtime + self.pev.cschargingtime + rp_dt) * UNITtimecost + self.pev.homechargingcost)
reward = -1 * self.pev.totalcost
else:
print('error final destination')
input()
return np.zeros((1, self.state_size)), -1, reward, done
else:
done = 1
reward = -50
return np.zeros((1, self.state_size)), -1, reward, done