def initialize_binapprox(problem_definition):
t = problem_definition["t"]
b_rel = problem_definition["b_rel"]
binapprox_init_optional_args = {}
try:
binapprox_init_optional_args["binary_threshold"] = \
problem_definition["binary_threshold"]
except KeyError:
pass
try:
binapprox_init_optional_args["off_state_included"] = \
problem_definition["off_state_included"]
except KeyError:
pass
try:
binapprox_init_optional_args["reduce_problem_size_before_solve"] = \
problem_definition["reduce_problem_size_before_solve"]
except KeyError:
pass
try:
binapprox = BinApprox(t=t, b_rel=b_rel, **binapprox_init_optional_args)
except Exception as e:
raise InvalidUsage(str(e), status_code=500)
return binapprox
class CombinaTestSingleInput(object):
b_rel = np.array([
9.164991131439303e-08, 9.164989910069793e-08, 9.164986964110127e-08,
9.164981439200778e-08, 9.164971954974675e-08, 9.164956322584703e-08,
9.164931096870898e-08, 9.164890871558705e-08, 9.16482717331077e-08,
9.164726728376491e-08, 9.164568747185336e-08, 9.164320671195729e-08,
9.163931511370584e-08, 9.163321414037958e-08, 9.162365315758747e-08,
9.160867333705047e-08, 9.158520628111474e-08, 9.154844464973005e-08,
9.149085452908308e-08, 9.140062376048531e-08, 9.125921952032561e-08,
9.103753238327347e-08, 9.06897600760712e-08, 9.01436326543539e-08,
8.928459967910412e-08, 8.792973591727924e-08, 8.578323033799388e-08,
8.235606307513253e-08, 7.680620700237228e-08, 6.75611734322906e-08,
5.110029321633913e-08, 4.403564664444702e-08, 4.106387808317555e-08,
3.983950373134124e-08, 3.934055898011604e-08, 3.9137914036044536e-08,
3.9054919200406776e-08, 3.901881246871395e-08, 3.899786135210769e-08,
3.897363059366587e-08, 3.8924779784938846e-08, 3.880822297418e-08,
3.852080876236079e-08, 3.780600129669751e-08, 3.601141047402357e-08,
3.139256276666938e-08, 2.9834455518510574e-08, 2.9318934446262996e-08,
2.914964834988244e-08, 2.9094176453158642e-08, 2.907592834306512e-08,
2.9069660025366436e-08, 2.906669316762665e-08, 2.906291807927123e-08,
2.9053105209808462e-08, 2.9023659839179072e-08, 2.893375174707503e-08,
2.865832883983718e-08, 2.781057309571279e-08, 2.516081216626174e-08,
1.6349821111021932e-08, 1.4641381658496285e-08, 1.4321695524188809e-08,
1.4262451136479392e-08, 1.4251493200337602e-08, 1.4249467289977935e-08,
1.424909202859219e-08, 1.4249016931279371e-08, 1.4248970149529801e-08,
1.4248781586888953e-08, 1.4247771634743814e-08, 1.4242308735857878e-08,
1.4212753082407813e-08, 1.4052702623315826e-08, 1.3181284088570942e-08,
8.268584499081758e-09, 7.688662077381325e-09, 7.621735475505242e-09,
7.61405669269621e-09, 7.613347382291173e-09, 7.614772557847257e-09,
7.628032105221415e-09, 7.743336673082607e-09, 7.774748608050492e-09,
7.781914757778366e-09, 7.777476689145606e-09, 7.768965238198726e-09,
7.809086255912332e-09, 8.347837492600157e-09, 1.3149458138366266e-08,
5.812677816280237e-08, 7.980857140061672e-08, 9.34899928315103e-08,
1.0296318896151463e-07, 1.0994080615953228e-07, 1.1547277950297847e-07,
1.2038557045733308e-07, 1.2548706350691847e-07, 1.3170304260698727e-07,
1.4020750175014196e-07, 1.526048438509999e-07, 1.7130057227378714e-07,
2.005178071994339e-07, 2.499858770490773e-07, 3.5728576775753295e-07,
0.20636133458659084, 0.206361456820869, 0.20636151047209217,
0.20636154006894097, 0.20636155692321392, 0.2063615660721068,
0.2063615709426096, 0.2063615747616593, 0.206361581065729,
0.20636159414396368, 0.2063616208814176, 0.20636168097818755,
0.2063617772714703, 0.2063618582371984, 0.20636189781122233,
0.20636190501221507, 0.2063619496037749, 0.2063620732923192,
0.20636249842130122, 0.2063773813145899, 0.252054864921568,
0.2520579982392464, 0.25288811587249843, 0.2593977742761429,
0.3340935108565165, 0.3340940442516569, 0.33409419385053196,
0.3340942499059316, 0.3340942652728998, 0.33409424511989244,
0.3135810041373855, 0.3135809645281175, 0.31358095939038705,
0.3135809756716781, 0.3135810190249818, 0.31358111927309246,
0.3135814262624077, 0.999999950926087, 0.999999974046433,
0.9999999783613379, 0.9999999789321505, 0.9999999787698426,
0.9999999779052294, 0.9999999747514201, 0.9999999550265595,
0.06453461127561795, 0.06453456953472514, 0.06453456775738621,
0.0645345926266459, 0.06453465796048136, 0.06453481999226998,
0.06453538382656564, 0.2131890930121317, 0.2131936809163467,
0.21319408556528727, 0.21319422088078965, 0.21319427911618313,
0.21319430204287756, 0.2131943004833132, 0.213194265080231,
2.287219254968173e-08, 1.0681031150093401e-08, 9.359002315937822e-09,
9.461200350554808e-09, 9.80570097938495e-09, 1.0901115627326692e-08,
1.2859087591404522e-08, 1.5439876067722526e-08, 1.881147846187437e-08,
2.325425317922466e-08, 2.905700927723151e-08, 3.620161250971683e-08,
4.3709179595637425e-08, 4.8807602615609486e-08, 4.614727473289024e-08,
2.319637525384318e-08, 2.3113256911779592e-08, 3.062367109037491e-08,
4.5803905719390506e-08, 7.373574744184745e-08, 1.249404286368601e-07,
2.24798134068856e-07, 4.5731586084303404e-07, 1.3208173664524378e-06,
0.0433240937001994, 0.04332605093577369, 0.043326318185106925,
0.04332641234938346, 0.043326451830150416, 0.04332646332568629,
0.04332645064755273, 0.043326464881207304, 0.04332652140919314,
0.9999999444789713, 0.9999999847239209, 0.9999999852314699,
0.9999999823896432, 0.9999999805987997, 0.9999999804592423,
0.9999999816949504, 0.9999999837988184, 0.9999999862895442,
0.999999988820378, 0.9999999911927993, 0.9999999933222256,
0.9999999952099605, 0.9999999968647432, 0.9999999982733226,
0.9999999994748296, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0,
1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0,
1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0,
1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0,
1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0,
1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0,
0.9999999976933017, 0.9999999927247275, 0.9999999813465971,
0.9999999516982475, 0.9999998765776738, 0.9999996930339815,
0.999999381480724, 0.9999952329064322, 0.33837190283783086,
0.33836939449774384, 0.33836820191669553, 0.3383674235415369,
0.3383668483297875, 0.3383663939468469, 0.33836601976451697,
0.3383657027491601, 0.3383654285764394, 0.33836518772741186,
0.33836497355245754, 0.3383647812216936, 0.3383646071134709,
0.33836444845465696, 0.3383643046158953, 0.33836417338226865,
0.338364053014526, 0.3383639421153582, 0.33836383954553606,
0.3383637443641662, 0.338363655785219, 0.33836357314532267,
0.33836349587948944, 0.3383634235025246, 0.3383633555945778,
0.3383632917897552, 0.3383632317670245, 0.3383631752430078,
0.3383631219726682, 0.33836307197017373, 0.33836302498944143,
0.3383629808173723, 0.3383629392632571, 0.33836290015502957,
0.33836286333638244, 0.3383628286684003, 0.3383627960634522,
0.3383627654812393, 0.33836273685899004, 0.33836271012357955,
0.3383626852031968, 0.33836266203107024, 0.33836264055345605,
0.33836262073384066, 0.33836260260332357, 0.33836258610656866,
0.33836257119446206, 0.33836255782361746, 0.338362545955937,
0.3383625355582329, 0.3383625266019055, 0.3383625190626729,
0.3383625129203459, 0.33836250815864527, 0.3383625047650565,
0.338362502730721, 0.33836250205035523
])
b_rel = np.vstack([b_rel, 1 - b_rel])
T = 240 * np.arange(0, b_rel.shape[1] + 1)
n_max_switches = [4, T.size]
binapprox = BinApprox(T, b_rel, binary_threshold=1e-3)
binapprox.set_n_max_switches(n_max_switches)
def test_check_binary_solution(self):
b_bin = self.binapprox.b_bin
assert_array_equal(np.squeeze(b_bin[0, :]), self.b_bin_check)
def test_check_objective(self):
b_bin = np.squeeze(self.binapprox.b_bin[0, :])
eta = self.binapprox.eta
Tg = self.T[1:] - self.T[:-1]
eta_check = 0
for k, b_bin_k in enumerate(b_bin):
eta_check_k = abs(
sum([Tg[j] * (self.b_rel[0][j] - b_bin[j]) for j in range(k)]))
if eta_check_k > eta_check:
eta_check = eta_check_k
self.assertAlmostEqual(eta, eta_check, 6)
def test_check_n_max_switches(self):
b_bin = np.squeeze(self.binapprox.b_bin[0, :])
n_switches = np.sum(np.absolute(b_bin[1:] - b_bin[:-1]))
self.assertTrue(n_switches <= self.n_max_switches[0])
0.3383634235025246, 0.3383633555945778, 0.3383632917897552,
0.3383632317670245, 0.3383631752430078, 0.3383631219726682,
0.33836307197017373, 0.33836302498944143, 0.3383629808173723,
0.3383629392632571, 0.33836290015502957, 0.33836286333638244,
0.3383628286684003, 0.3383627960634522, 0.3383627654812393,
0.33836273685899004, 0.33836271012357955, 0.3383626852031968,
0.33836266203107024, 0.33836264055345605, 0.33836262073384066,
0.33836260260332357, 0.33836258610656866, 0.33836257119446206,
0.33836255782361746, 0.338362545955937, 0.3383625355582329,
0.3383625266019055, 0.3383625190626729, 0.3383625129203459,
0.33836250815864527, 0.3383625047650565, 0.338362502730721,
0.33836250205035523
])
t = 240 * pl.arange(0, b_rel.size + 1)
binapprox = BinApprox(t = t, b_rel = b_rel, binary_threshold = 1e-3, \
off_state_included = False)
sur = CombinaSUR(binapprox)
sur.solve()
b_bin = binapprox.b_bin
pl.figure()
pl.step(t[:-1],
b_rel,
label="b_rel",
color="C0",
linestyle="dashed",
where="post")
pl.step(t[:-1], b_bin.T, label="b_bin", color="C0", where="post")
pl.xlabel("t")
pl.ylabel("b")
import pylab as pl
from pycombina import BinApprox, CombinaBnB, CombinaMILP, CombinaSUR
pl.close("all")
data = pl.loadtxt("data/mmlotka_nt_12000_400.csv", delimiter=" ", skiprows=1)
dN = 8
t = data[::dN, 0]
b_rel = data[:-1:dN, 3:]
max_switches = [5, 2, 3]
binapprox = BinApprox(t = t, b_rel = b_rel, binary_threshold = 1e-3, \
off_state_included = True)
binapprox.set_n_max_switches(n_max_switches=max_switches)
#binapprox.set_valid_controls_for_interval((0, 2), [1,0,0])
#binapprox.set_valid_control_transitions(0, [1,0,1])
#binapprox.set_min_up_times([2.0, 2.0, 2.0])
#binapprox.set_cia_norm("row_sum_norm")
#binapprox.set_b_bin_pre([1,0,0])
combina = CombinaBnB(binapprox)
combina.solve(
use_warm_start=False) #, gurobi_opts = {"TimeLimit": 20, "MIPGap": 0.4})
b_bin_orig = pl.asarray(binapprox.b_bin)
0.3383629392632571, 0.33836290015502957, 0.33836286333638244,
0.3383628286684003, 0.3383627960634522, 0.3383627654812393,
0.33836273685899004, 0.33836271012357955, 0.3383626852031968,
0.33836266203107024, 0.33836264055345605, 0.33836262073384066,
0.33836260260332357, 0.33836258610656866, 0.33836257119446206,
0.33836255782361746, 0.338362545955937, 0.3383625355582329,
0.3383625266019055, 0.3383625190626729, 0.3383625129203459,
0.33836250815864527, 0.3383625047650565, 0.338362502730721,
0.33836250205035523
])
b_rel = pl.vstack([b_rel, 1 - b_rel])
t = 240 * pl.arange(0, b_rel.shape[1] + 1)
binapprox = BinApprox(t=t, b_rel=b_rel, binary_threshold=1e-3)
sur = CombinaSUR(binapprox)
sur.solve()
b_bin = binapprox.b_bin
pl.figure()
pl.step(t[:-1],
b_rel[0, :],
label="b_rel",
color="C0",
linestyle="dashed",
where="post")
pl.step(t[:-1], b_bin[0, :], label="b_bin", color="C0", where="post")
pl.xlabel("t")
pl.ylabel("b")