def test_f(self):
'''
'''
cases = {}
# Each case: t, N, Dmax, Smax, c_h, adaptive, slope, placement_optimistic, bcharge, pt = result
# Case 1a: consumer first, thus generator last, who creates negative 1
# consumer always adds 2.75
cases[(26, 10, 3, 1, 1.2, True, .5, True, 0, .5)] = (round(5 * 2.787 + 5 * -1, 4), -1)
# Case 1b: Now price is adapted, so they buy 2.887
cases[(26, 10, 3, 1, 1.2, True, .5, True, 0, .3)] = (round(5 * 2.887 + 5 * -1, 4), -1)
# Case 1c: Now try pessimistic
cases[(26, 10, 3, 1, 1.2, True, .5, False, 0, .3)] = (round(5 * 2.887 + 5 * -1, 4), -5)
# Case 2: try higher N
cases[(26, 30, 3, 1, 1.2, True, .5, True, 0, .3)] = (round(15 * 2.887 + 15 * -1, 4), -1)
# Case 3: test non-adaptive case. They buy 2.924
cases[(26, 10, 3, 1, 1.2, False, .5, True, 0, .3)] = (round(5 * 2.924 + 5 * -1, 4), -1)
# Case 4: different time step: let house batteries charge, no PV
cases[(38, 10, 3, 1, 1.2, True, .5, True, 0, .3)] = (round(5 * 3.887, 4), 0)
# Case 5a: our battery charges 0.6
cases[(26, 10, 3, 1, 1.2, True, .5, True, 0.6, .3)] = \
(round(.6 + (5 * 2.887 + 5 * -1), 4), -1 + 0.6)
# Case 5b: and discharge
cases[(26, 10, 3, 1, 1.2, True, .5, True, -0.6, .3)] = \
(round(-.6 + (5 * 2.887 + 5 * -1), 4), -1 - 0.6)
for c in cases.keys():
T = c[0]
s = Street(T=T+1, N=c[1], C=None, Dmax=c[2], Smax=c[3], pmax=.452, pmin=.074, c_h=c[4],
adaptive=c[5], slope=c[6], placement_optimistic=c[7])
#s.draw(T, c[8], c[9])
fp, fm = s.f(T, c[8], c[9])
self.assertEqual((round(fp, 4), round(fm, 4)), cases[c])
def main(conf, log, seed):
ws = [.05, .2, .5, 1, 2]
battery_types = (NoBattery, OfflineBattery, BaseBattery,
HeuristicBattery, NonDeterministicOfflineBattery)
#battery_types = (NoBattery, BaseBattery, HeuristicBattery,)
T = 48 # this is given by the price data structure: half-hour intervals
N = conf.getint('params', 'N')
C = conf.getint('params', 'C')
B = conf.getfloat('params', 'B')
alpha = conf.getfloat('params', 'alpha')
max_rate = conf.getfloat('params', 'max_rate')
econ_slope = conf.getfloat('params', 'econ_slope')
Dmax = conf.getfloat('params', 'Dmax')
Smax = conf.getfloat('params', 'Smax')
pmax = conf.getfloat('params', 'pmax')
pmin = conf.getfloat('params', 'pmin')
c_h = conf.getfloat('params', 'c_h')
adaptive = conf.getboolean('params', 'adaptive')
placement_optimistic = conf.getboolean('params', 'placement_optimistic')
LP_max_runtime = conf.getint('params', 'LP_max_runtime') # in minutes
LP_max_k = conf.getint('params', 'LP_max_k') # max. assumed k
debug = conf.getboolean('params', 'debug')
# get price series for us to use in this run
pfile = open('price_data/APXReferencePriceDataUkPowerMarket2012_noweekends.csv', 'r')
random.seed(seed)
row = random.randint(1, 219) # 219: number of days in data
for _ in range(row):
pfile.readline()
line = pfile.readline().strip().split('\t')
day = line[0]
actual_prices = [float(p) / 300. for p in line[1:]]
# get expected prices for the right month from pre-computed avg file
exp_prices_in = open('price_data/APXReferencePriceDataUkPowerMarket2012_noweekends_avg.csv', 'r')
exp_prices_in.readline() # Title
for line in exp_prices_in.readlines():
line = line.strip().split('\t')
if '-{}-'.format(line[0]) in day: # we reached the correct month
avg_price = float(line[1])
exp_prices = [float(p) for p in line[2:]]
break
street = Street(T=T, N=N, C=C, Dmax=Dmax, Smax=Smax, pmax=pmax, pmin=pmin,
c_h=c_h, adaptive=adaptive, slope=econ_slope,
placement_optimistic=placement_optimistic)
log.write('# w, ')
for btype in battery_types:
log.write('{}, '.format(btype.name))
log.write('\n')
for w in ws:
line = '{}, '.format(w)
for btype in battery_types:
b = btype(capacity=B, efficiency=alpha, max_rate=max_rate,
street=street, exp_prices=exp_prices, avg_price=avg_price,
T=T, c_h=c_h)
if btype in (OfflineBattery, DeterministicH2Battery):
b.exp_prices = actual_prices
if btype in (OfflineBattery, NonDeterministicOfflineBattery):
b.max_runtime = LP_max_runtime
b.max_k = LP_max_k
b.solve_offline(w)
accumulated_costs = 0
for t in xrange(0, T):
pt = actual_prices[t]
bcharge = b.compute_charge(t, pt)
if debug:
fp, fm = street.f(t, bcharge, pt)
print("[{}][{}] exp. price: {}, act. price:{}, fp:{}, fm:{}, magnitude before:{}, magnitude after:{}, b.level:{}, bcharge:{}".format(
t, b.name, round(exp_prices[t], 2),
round(pt, 2), round(fp, 2), round(fm, 2),
round(street.maxf(t, 0, pt), 2),
round(street.maxf(t, bcharge, pt), 2),
round(b.level, 2),
round(bcharge, 2)))
revenue = b.execute_charge(bcharge, pt)
cost = street.cost(t, bcharge, pt)
accumulated_costs += w * cost - revenue
# now sell the rest in the battery at an assumed worth of the
# average price of that month
if not btype is NoBattery:
accumulated_costs -= b.level * avg_price
b.level = 0
line += '{},'.format(accumulated_costs)
if debug:
print "Acc. Costs: ", accumulated_costs
log.write('{}\n'.format(line[:-1])) # write line w/o last comma
log.flush()
log.close()
