def __init__(self, num_stages = 12,\
):
self.num_stages = num_stages
self.stations = Stations()
self.action_L = self.stations.pmin
self.action_H = self.stations.pmax
self.action_space = abox.ABox(self.action_L,
self.action_H,
dtype=np.float64)
self.observation_space = mbox.MBox(self.stations.num_cars,
self.stations.num_stations)
#self.observation= np.multiply(np.ones(self.stations.num_stations), self.stations.num_cars/self.num_stations)
self.t = 0
def __init__(self,
num_stations=N_def,
num_cars=MAX_CARS_def,
min_price=pmin_def,
travel_time_btwn_stat=T_def,
a=a_def,
b=b_def):
self.__version__ = "0.1.0"
logging.info("CarsharingEnv - Version {}".format(self.__version__))
# General variables defining the environment
self.N = num_stations # Number of stations
self.MAX_CARS = num_cars # Max number of cars in the system
self.set_pmin = min_price
# self.T = np.array([ ])
# self.T=np.random.randint(1,4, size=(self.N, self.N))
self.T = travel_time_btwn_stat
self.kmax = np.max(self.T)
self.a = a
self.b = b
self.pmin = min_price
self.pmax = np.empty(self.N)
self.pmax = self.a / (
-1 * self.b
) - 0.1 # -1 set the price such that the expected demand cannot be negative
self.pmax = np.around(self.pmax, 1)
self.dmin = self.D(self.pmax)
self.dmax = self.D(self.pmin)
self.Nik = []
for i in range(self.N):
self.Nik.append(
[np.where(self.T[:, i] == k) for k in range(1, self.kmax + 1)])
self.action_space = spaces.Box(self.pmin, self.pmax, dtype=np.float32)
self.s = spaces.Box(0,
self.kmax,
shape=(self.N, self.kmax),
dtype=np.uint8)
self.x = mbox.MBox(self.MAX_CARS, self.N)
self.observation_space = spaces.Tuple((self.x, self.s))
self.t = 0
def __init__(self,
num_stations=N_def,
num_cars=MAX_CARS_def,
aa_min_price=aa_pmin_def,
aa_max_price=aa_pmax_def,
ab_min_price=ab_pmin_def,
ab_max_price=ab_pmax_def,
aa=aa_def,
ab=ab_def,
bb=bb_def,
ba=ba_def,
discount_rate=discount_rate_def,
num_stages=num_stages_def,
aa_epsSupL=aa_epsSupL_def,
aa_epsSupH=aa_epsSupH_def,
ab_epsSupL=ab_epsSupL_def,
ab_epsSupH=ab_epsSupH_def):
self.__version__ = "0.1.0"
logging.info("CarsharingEnv - Version {}".format(self.__version__))
self.N = num_stations # Number of stations
self.MAX_CARS = num_cars # Max number of cars in the system
#return trip demand model
self.aa = aa
self.bb = bb
#one way trip demand model
self.ab = ab
self.ba = ba
#return trips epsilon support
self.aa_epsSupL = aa_epsSupL
self.aa_epsSupH = aa_epsSupH
#one trips epsilon support
self.ab_epsSupL = ab_epsSupL
self.ab_epsSupH = ab_epsSupH
#return min/max price
self.aa_pmin = aa_min_price
self.aa_pmax = np.minimum(
np.around((self.aa + self.aa_epsSupL) / (-1 * self.bb), 1),
aa_max_price)
#return fixed demand and price for one-way station
self.aa_p = (self.aa_pmin + self.aa_pmax) * 0.5
self.aa_d = self.DAA(self.aa_p)
#one_way trips min/max price
self.ab_pmin = ab_min_price
self.ab_pmax = np.minimum(
np.around((self.ab + self.ab_epsSupL) / (-1 * self.ba), 1),
ab_max_price)
self.aa_dmin = self.DAA(self.aa_pmax)
self.aa_dmax = self.DAA(self.aa_pmin)
self.ab_dmin = self.DAB(self.ab_pmax)
self.ab_dmax = self.DAB(self.ab_pmin)
#for the actions it is better to separate AA and A_all_other, since AAs are fixed
# theta variables !!
#easy need to create a variable
self.action_L = np.concatenate((self.ab_dmin, np.zeros(self.N)))
self.action_H = np.concatenate(
(self.ab_dmax, np.multiply(np.ones(self.N), self.MAX_CARS)))
self.action_space = abox.ABox(self.action_L,
self.action_H,
dtype=np.float32)
# self.s=spaces.Box(0, self.kmax,shape=(self.N,self.kmax), dtype=np.uint8)
# self.observation_space = spaces.Tuple((self.x,self.s))
self.observation = np.multiply(
np.ones(self.N),
self.MAX_CARS / self.N) #mbox.randomize(self.MAX_CARS, self.N)
self.observation_space = mbox.MBox(self.MAX_CARS, self.N)
self.t = 0