def test_return_embedding(self):
nodelist = [0, 1, 2]
edgelist = [(0, 1), (1, 2), (0, 2)]
sampler = EmbeddingComposite(
dimod.StructureComposite(dimod.NullSampler(), nodelist, edgelist))
sampleset = sampler.sample_ising({'a': -1}, {'ac': 1},
return_embedding=True)
self.assertIn('embedding', sampleset.info['embedding_context'])
embedding = sampleset.info['embedding_context']['embedding']
self.assertEqual(set(embedding), {'a', 'c'})
self.assertIn('chain_break_method',
sampleset.info['embedding_context'])
self.assertEqual(
sampleset.info['embedding_context']['chain_break_method'],
'majority_vote') # the default
self.assertIn('embedding_parameters',
sampleset.info['embedding_context'])
self.assertEqual(
sampleset.info['embedding_context']['embedding_parameters'],
{}) # the default
self.assertIn('chain_strength', sampleset.info['embedding_context'])
self.assertEqual(sampleset.info['embedding_context']['chain_strength'],
1.0) # the default
# default False
sampleset = sampler.sample_ising({'a': -1}, {'ac': 1})
self.assertNotIn('embedding_context', sampleset.info)
def test_return_embedding_as_class_variable(self):
nodelist = [0, 1, 2]
edgelist = [(0, 1), (1, 2), (0, 2)]
sampler = EmbeddingComposite(
dimod.StructureComposite(dimod.NullSampler(), nodelist, edgelist))
# temporarily change return_embedding_default
EmbeddingComposite.return_embedding_default = True
sampleset = sampler.sample_ising({'a': -1}, {'ac': 1})
self.assertIn('embedding', sampleset.info['embedding_context'])
embedding = sampleset.info['embedding_context']['embedding']
self.assertEqual(set(embedding), {'a', 'c'})
self.assertIn('chain_break_method',
sampleset.info['embedding_context'])
self.assertEqual(
sampleset.info['embedding_context']['chain_break_method'],
'majority_vote') # the default
self.assertIn('embedding_parameters',
sampleset.info['embedding_context'])
self.assertEqual(
sampleset.info['embedding_context']['embedding_parameters'],
{}) # the default
self.assertIn('chain_strength', sampleset.info['embedding_context'])
self.assertEqual(sampleset.info['embedding_context']['chain_strength'],
1.414) # the default
# restore the default
EmbeddingComposite.return_embedding_default = False
def test_instantiation(self):
for factory in [dimod.ExactSolver, dimod.RandomSampler, dimod.SimulatedAnnealingSampler]:
sampler = dimod.StructureComposite(factory(), [0, 1, 2], [(0, 1), (0, 2), (1, 2)])
dit.assert_sampler_api(sampler)
dit.assert_composite_api(sampler)
dit.assert_structured_api(sampler)
def test_minimize_energy_chain_break_method(self):
# bug report https://github.com/dwavesystems/dwave-system/issues/206
Q = {
(0, 1): 1,
(0, 2): 1,
(0, 3): 1,
(1, 2): 1,
(1, 3): 1,
(2, 3): 1,
(0, 0): 1,
(1, 1): 1,
(2, 2): 1,
(3, 3): 1
}
S = {(0, 1): 1, (0, 2): 1, (0, 3): 1, (1, 2): 1, (1, 3): 1, (2, 3): 1}
bqm = dimod.BinaryQuadraticModel.from_qubo(Q)
G = dnx.chimera_graph(4)
sampler = dimod.StructureComposite(dimod.ExactSolver(), G.nodes,
G.edges)
embedding = {0: [55], 1: [48], 2: [50, 53], 3: [52, 51]}
composite = FixedEmbeddingComposite(sampler, embedding)
cbm = chain_breaks.MinimizeEnergy(bqm, embedding)
composite.sample(bqm, chain_break_method=cbm).resolve()
def test_smoke(self):
nodelist = [0, 1, 2, 3]
edgelist = [(0, 1), (1, 2), (2, 3), (0, 3)]
child = dimod.StructureComposite(dimod.NullSampler(), nodelist, edgelist)
sampler = AutoEmbeddingComposite(child)
sampler.sample_ising({}, {(0, 1): -1})
sampler.sample_ising({}, {('a', 0): -1})
def test_find_embedding_kwarg(self):
child = dimod.StructureComposite(dimod.NullSampler(), [0, 1], [(0, 1)])
def my_find_embedding(S, T):
# does nothing
return {v: [v] for v in set().union(*S)}
sampler = EmbeddingComposite(child, find_embedding=my_find_embedding)
# nothing breaks
sampler.sample_ising({0: -1}, {})
def test_embedding_parameters_sample(self):
child = dimod.StructureComposite(dimod.NullSampler(), [0, 1], [(0, 1)])
def my_find_embedding(S, T, a):
assert a == -1
return {v: [v] for v in set().union(*S)}
sampler = EmbeddingComposite(child, find_embedding=my_find_embedding)
# nothing breaks
sampler.sample_ising({0: -1}, {}, embedding_parameters={'a': -1})
def test_warnings(self):
G = dnx.chimera_graph(12)
sampler = EmbeddingComposite(
dimod.StructureComposite(dimod.RandomSampler(), G.nodes, G.edges))
# this will need chains lengths > 7
J = {uv: -1 for uv in itertools.combinations(range(40), 2)}
ss = sampler.sample_ising({}, J, warnings='SAVE')
self.assertIn('warnings', ss.info)
def test_composed_sampler(self):
nodelist = list(range(5))
edgelist = list(itertools.combinations(nodelist, 2))
structured_sampler = dimod.StructureComposite(dimod.NullSampler(),
nodelist, edgelist)
sampler = dimod.TrackingComposite(structured_sampler)
structure = dimod.child_structure_dfs(sampler)
self.assertEqual(structure.nodelist, nodelist)
self.assertEqual(structure.edgelist, edgelist)
def test_instantiation(self):
C = dnx.chimera_graph(2, 2, 4)
for factory in [
dimod.ExactSolver, dimod.RandomSampler,
dimod.SimulatedAnnealingSampler
]:
structsampler = dimod.StructureComposite(factory(),
nodelist=C.nodes(),
edgelist=C.edges())
sampler = CheckerboardTransformComposite(structsampler, C)
dit.assert_sampler_api(sampler)
dit.assert_composite_api(sampler)
def test_broken_find_embedding(self):
nodelist = [0, 1, 2, 3]
edgelist = [(0, 1), (1, 2), (2, 3), (0, 3)]
child = dimod.StructureComposite(dimod.NullSampler(), nodelist, edgelist)
def find_embedding(*args, **kwargs):
raise NotImplementedError
sampler = AutoEmbeddingComposite(child, find_embedding=find_embedding)
sampler.sample_ising({}, {(0, 1): -1})
with self.assertRaises(NotImplementedError):
sampler.sample_ising({}, {('a', 0): -1})
def test_external_embedding_sampler(self):
bqm = dimod.BinaryQuadraticModel.from_ising({'a': 1}, {})
init = State.from_subproblem(bqm, embedding={'a': [0]})
sampler = dimod.StructureComposite(
SimulatedAnnealingSampler(), nodelist=[0], edgelist=[])
workflow = QPUSubproblemExternalEmbeddingSampler(qpu_sampler=sampler)
# run mock sampling
res = workflow.run(init).result()
# verify mock sampler received custom kwargs
self.assertEqual(res.subsamples.first.energy, -1)
def example_instance():
nP = 2
jobs = [1, 2, 3, 4]
diff = []
embedding = {
0: [1760, 1632, 1764],
1: [1628, 1636, 1634],
2: [1638, 1630, 1627],
3: [1752, 1624, 1759, 1767],
4: [1763, 1635, 1765],
5: [1761, 1633],
6: [1754, 1626, 1758, 1766],
7: [1631, 1639, 1625],
8: [1637, 1629]
}
# base_sampler = neal.SimulatedAnnealingSampler()
base_sampler = DWaveSampler()
#base_sampler.properties['extended_j_range'] = [-10.0, 10.0]
G = dnx.chimera_graph(16, 16, 4)
nodelist = G.nodes()
edgelist = G.edges()
sampler = dimod.StructureComposite(base_sampler, nodelist, edgelist)
new_sampler = FixedEmbeddingComposite(sampler, embedding)
# print new_sampler.properties
# print new_sampler.parameters
#new_sampler = VirtualGraphComposite(sampler, embedding, chain_strength=-6.0)
Q = create.createQUBO(jobs, nP, diff)
solutions = {
'energy': [],
"valid": [],
'max_time': [],
'chain_break_fraction': []
}
for k in range(10):
s = solve.solveQUBO(Q, jobs, nP, new_sampler, 0)
solutions['energy'].append(s['energy'])
solutions['valid'].append(s['valid'])
solutions['max_time'].append(s['max_time'])
solutions['chain_break_fraction'].append(s['chain_break_fraction'])
solve.evaluate(solutions)
print(solutions)
print("Durchschnittliche Anteil gebrochender Ketten: ",
sum(solutions['chain_break_fraction']) / 100)
def test_warning_chain_strength(self):
G = dnx.chimera_graph(12)
sampler = EmbeddingComposite(
dimod.StructureComposite(dimod.RandomSampler(), G.nodes, G.edges))
J = {(0, 1): 100}
ss = sampler.sample_ising({}, J, warnings='SAVE')
self.assertIn('warnings', ss.info)
count = 0
for warning in ss.info['warnings']:
if issubclass(warning['type'], ChainStrengthWarning):
count += 1
self.assertEqual(count, 1)
def test_chimera(self):
C = dnx.chimera_graph(4)
nodelist = list(C.nodes())
edgelist = list(C.edges())
structsampler = dimod.StructureComposite(dimod.RandomSampler(),
nodelist=nodelist,
edgelist=edgelist)
Q = {(v, v): 0.1 for v in nodelist}
Q.update({edge: -1.0 for edge in edgelist})
sampler = CheckerboardTransformComposite(structsampler, C)
response = sampler.sample_qubo(Q, num_reads=1000)
dit.assert_response_energies(response,
dimod.BinaryQuadraticModel.from_qubo(Q))
def test_warnings_as_class_variable(self):
G = dnx.chimera_graph(12)
sampler = EmbeddingComposite(
dimod.StructureComposite(dimod.RandomSampler(), G.nodes, G.edges))
# this will need chains lengths > 7
J = {uv: -1 for uv in itertools.combinations(range(40), 2)}
EmbeddingComposite.warnings_default = 'SAVE'
ss = sampler.sample_ising({}, J)
self.assertIn('warnings', ss.info)
EmbeddingComposite.warnings_default = 'IGNORE' # restore default
def test_keyer(self):
C4 = dnx.chimera_graph(4)
nodelist = sorted(C4.nodes)
edgelist = sorted(sorted(edge) for edge in C4.edges)
child = dimod.StructureComposite(dimod.NullSampler(), nodelist, edgelist)
embedding = {0: [49, 53], 1: [52], 2: [50]}
sampler = FixedEmbeddingComposite(child, embedding)
with self.assertRaises(dwave.embedding.exceptions.MissingChainError):
sampler.sample_qubo({(1, 4): 1})
sampler.sample_qubo({(1, 2): 1}).record # sample and resolve future
sampler.sample_qubo({(1, 1): 1}).record # sample and resolve future
def test_keyerror(self):
C16 = dnx.chimera_graph(16)
child_sampler = dimod.RandomSampler()
nodelist = sorted(C16.nodes)
edgelist = sorted(sorted(edge) for edge in C16.edges)
struc_sampler = dimod.StructureComposite(child_sampler, nodelist,
edgelist)
embedding = {0: [391, 386], 1: [390], 2: [385]}
sampler = FixedEmbeddingComposite(struc_sampler, embedding)
with self.assertRaises(ValueError):
sampler.sample_qubo({(1, 4): 1})
sampler.sample_qubo({(1, 2): 1}).record # sample and resolve future
sampler.sample_qubo({(1, 1): 1}).record # sample and resolve future
def test_return_embedding(self):
nodelist = [0, 1, 2]
edgelist = [(0, 1), (1, 2), (0, 2)]
sampler = EmbeddingComposite(
dimod.StructureComposite(dimod.NullSampler(), nodelist, edgelist))
sampleset = sampler.sample_ising({'a': -1}, {'ac': 1},
return_embedding=True)
self.assertIn('embedding', sampleset.info)
embedding = sampleset.info['embedding']
self.assertEqual(set(embedding), {'a', 'c'})
# default False
sampleset = sampler.sample_ising({'a': -1}, {'ac': 1})
self.assertNotIn('embedding', sampleset.info)
def test_scale_aware_scale_composite(self):
nodelist = [0, 1, 2]
edgelist = [(0, 1), (1, 2), (0, 2)]
embedding = {'a': [0], 'b': [1, 2]}
sampler = FixedEmbeddingComposite(dimod.TrackingComposite(
ScaleComposite(
dimod.StructureComposite(dimod.NullSampler(), nodelist,
edgelist))),
embedding=embedding,
scale_aware=True)
_ = sampler.sample_ising({'a': 100, 'b': -100}, {'ab': 300})
self.assertIn('ignored_interactions', sampler.child.input)
ignored = sampler.child.input['ignored_interactions']
self.assertTrue(ignored == [(1, 2)] or ignored == [(2, 1)])
def test_transforms_exact(self):
C = dnx.chimera_graph(2, 2, 2)
nodelist = list(C.nodes())
edgelist = list(C.edges())
structsampler = dimod.StructureComposite(dimod.ExactSolver(),
nodelist=nodelist,
edgelist=edgelist)
sampler = CheckerboardTransformComposite(structsampler,
C,
aggregate=True)
h = {v: 0.1 for v in nodelist}
J = {edge: -1.0 for edge in edgelist}
response = sampler.sample_ising(h, J)
# All 4 gauges must return same samples
for datum in response.data():
self.assertEqual(datum.num_occurrences, 4)
dit.assert_response_energies(
response, dimod.BinaryQuadraticModel.from_ising(h, J))
def test_bug60(self):
for __ in range(100):
# make a small K5 bqm
bqm = dimod.BinaryQuadraticModel.empty(dimod.BINARY)
for v in range(5):
bqm.add_variable(v, random.uniform(-1, 1))
for u, v in itertools.combinations(range(5), 2):
bqm.add_interaction(u, v, random.uniform(-1, 1))
sampler_exact = dimod.ExactSolver()
# get the structure of a C1,2 chimera lattice
C12 = dnx.chimera_graph(2, 2, 3)
nodelist = sorted(C12)
edgelist = sorted(sorted(edge) for edge in C12.edges)
sampler_structured = dimod.StructureComposite(sampler_exact,
nodelist=nodelist,
edgelist=edgelist)
sampler_embedding = EmbeddingComposite(sampler_structured)
dtest.assert_sampler_api(sampler_embedding)
resp_exact = sampler_exact.sample(bqm)
resp_emb = sampler_embedding.sample(bqm)
for sample, energy in resp_emb.data(['sample', 'energy']):
self.assertEqual(bqm.energy(sample), energy)
ground_exact = dict(next(iter(resp_exact)))
ground_embed = dict(next(iter(resp_emb)))
self.assertEqual(ground_embed, ground_exact)
def test_transform_embedded(self):
C = dnx.chimera_graph(1)
nodelist = list(C.nodes())
edgelist = list(C.edges())
structsampler = dimod.StructureComposite(dimod.ExactSolver(),
nodelist=nodelist,
edgelist=edgelist)
gauges_sampler = CheckerboardTransformComposite(structsampler,
C,
aggregate=True)
sampler = FixedEmbeddingComposite(gauges_sampler, {
'a': [0, 4],
'b': [1, 5],
'c': [2, 6]
})
h = {'a': .5, 'c': 0}
J = {('a', 'c'): -1}
response = sampler.sample_ising(h, J)
# All 4 gauges must return same samples
for datum in response.data():
self.assertEqual(datum.num_occurrences, 4)
dit.assert_response_energies(
response, dimod.BinaryQuadraticModel.from_ising(h, J))
def composite_factory(sampler):
return dimod.StructureComposite(sampler, [0, 1, 2],
[(0, 1), (0, 2), (1, 2)])
def example_instance_delete_given(percentage):
nP = 3
jobs = [5, 7, 3, 5, 9, 6, 3, 1, 3, 1]
diff = [1, 2, 2, 5]
base_sampler = DWaveSampler()
nodelist = base_sampler.nodelist
edgelist = base_sampler.edgelist
# print new_sampler.properties
# print new_sampler.parameters
# new_sampler = VirtualGraphComposite(sampler, embedding, chain_strength=-6.0)
Q, num_entries, max_entry = create.createQUBO(jobs, nP, diff)
# nur noch nach absolutwerten sortiert, enthält nur die keys nicht die values
sorted_q = sorted(Q, key=lambda dict_key: abs(Q[dict_key]))
embedding = {
0: [
622, 614, 606, 598, 726, 594, 722, 978, 630, 638, 850, 593, 465,
337, 602, 1106, 590
],
1: [
1090, 834, 1110, 578, 1102, 1094, 962, 1118, 706, 1134, 1126, 1142,
1150, 710, 836, 718, 1112, 702, 583
],
2: [894, 830, 838, 846, 854, 886, 878, 870, 862, 851, 723, 595, 979],
3: [362, 106, 490, 234, 618, 749, 1002, 746, 874, 110, 118, 126, 1130],
4: [
342, 334, 326, 318, 382, 374, 569, 953, 366, 358, 350, 313, 441,
697, 825, 1081
],
5: [
315, 325, 341, 333, 317, 365, 357, 349, 443, 571, 699, 381, 827,
955, 373
],
6: [123, 635, 379, 1019, 763, 251, 891, 1147, 507, 1275, 1151],
7: [764, 756, 716, 748, 740, 732, 724, 708, 700],
8: [
591, 599, 607, 345, 985, 601, 729, 857, 615, 473, 639, 1113, 623,
631
],
9: [
511, 455, 704, 448, 471, 463, 960, 576, 479, 487, 496, 495, 503,
320, 624, 832, 1088
],
10: [
709, 1141, 971, 1099, 1109, 843, 717, 844, 715, 701, 1101, 1117,
1125, 1133, 725, 733, 1149, 741
],
11: [761, 1017, 633, 889, 505, 377, 249, 1145, 121, 1273, 127],
12: [893, 829, 885, 877, 837, 845, 853, 861, 869, 880, 752, 1008],
13: [1007, 999, 991, 959, 967, 975, 983, 1015, 1023, 977, 849, 721],
14: [
1127, 961, 1119, 833, 1095, 1103, 1111, 1089, 705, 839, 1135, 831,
847, 1120, 1248, 1254, 1262, 1270, 1278, 855, 863
],
15: [
608, 589, 605, 1098, 597, 613, 480, 352, 586, 842, 714, 970, 224,
96, 101, 109, 117, 125
],
16: [611, 355, 483, 227, 253, 245, 237, 229, 739, 867, 871, 995, 1123],
17: [491, 1003, 107, 363, 619, 747, 246, 238, 235, 875, 254, 1131],
18: [482, 758, 766, 750, 742, 994, 738, 610, 866, 1122, 354],
19: [
988, 835, 1004, 996, 980, 707, 964, 972, 992, 864, 868, 876, 884,
892, 956, 963
],
20: [632, 1144, 888, 504, 1016, 248, 376, 760, 1272, 120],
21: [
982, 958, 966, 974, 950, 946, 818, 690, 990, 998, 1006, 1014, 1022,
692
],
22: [
442, 452, 468, 460, 1082, 444, 570, 698, 826, 954, 476, 484, 458,
492, 500, 508, 314
],
23: [728, 216, 344, 856, 472, 236, 228, 220, 600, 244, 984, 734, 252],
24: [
510, 502, 494, 865, 486, 481, 478, 353, 609, 993, 737, 1121, 470,
462
],
25: [
84, 72, 456, 200, 328, 92, 76, 100, 98, 226, 968, 108, 116, 584,
461, 469, 712, 124, 840, 1096
],
26: [703, 751, 719, 727, 767, 759, 743, 735, 711, 695, 736],
27: [475, 493, 485, 477, 987, 603, 731, 859, 501, 509, 347, 1115],
28:
[1011, 629, 371, 621, 627, 637, 755, 243, 499, 883, 1139, 115, 1267],
29: [
1005, 997, 989, 1013, 965, 976, 973, 981, 957, 1021, 848, 720, 592,
596, 588
],
30:
[329, 343, 335, 585, 383, 359, 351, 457, 367, 375, 713, 841, 969, 338],
31: [
340, 332, 324, 364, 348, 356, 316, 312, 440, 568, 696, 824, 952,
372, 380, 1080
],
32: [
1116, 1114, 1124, 1084, 1092, 1100, 1108, 986, 730, 1132, 858,
1140, 1148, 1097, 1083
],
33: [1018, 762, 250, 634, 122, 506, 378, 890, 1020, 1146, 1274],
34: [757, 626, 1010, 754, 882, 498, 1138, 370, 242, 765, 114, 1266],
35:
[1001, 617, 636, 628, 620, 873, 489, 745, 361, 233, 105, 1129, 612],
36: [1009, 625, 881, 497, 369, 241, 895, 753, 887, 1137, 113, 1265],
37: [
255, 247, 872, 239, 488, 616, 232, 360, 744, 1000, 1128, 240, 112,
231
]
}
sampler = dimod.StructureComposite(base_sampler, nodelist, edgelist)
new_sampler = FixedEmbeddingComposite(sampler, embedding)
alldata = []
n_deleted = []
v_deleted = []
for p in percentage:
print("Percentage: ", p)
Q_deleted, max, deleted = get_given_percentage_qubo(
p, copy.deepcopy(Q), copy.deepcopy(sorted_q), num_entries)
n_deleted.append(deleted)
print("Größter gelöschter Wert: ", max)
v_deleted.append(max)
solution = []
kaputt = 0
for i in range(100):
result = solve.solveQUBO(Q_deleted, jobs, nP, new_sampler)
#print(result)
solution.append(result)
if result['valid'] == False:
kaputt += 1
print(kaputt, " kaputte Lösungen für Threshold und ", p)
alldata.append(solution)
dist = []
for i in range(len(alldata)):
dist.append(
[alldata[i][j]['max_time'] for j in range(len(alldata[i]))])
print(dist)
rc('font', **{'family': 'sans-serif', 'sans-serif': ['Helvetica']})
## for Palatino and other serif fonts use:
# rc('font',**{'family':'serif','serif':['Palatino']})
rc('text', usetex=True)
plt.rc('text', usetex=True)
plt.rc('font', family='serif')
plt.boxplot(dist)
plt.xticks([
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20,
21
], [
0.0, 0.05, 0.1, 0.15, 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5, 0.55, 0.6,
0.65, 0.7, 0.75, 0.8, 0.85, 0.9, 0.95, 1.0
])
plt.xlabel("Prozentanzahl entfernter Elemente")
plt.ylabel("Umsteigedistanzen")
plt.title("Jeweils 200 Aufrufe, beste Ergebnisse, Datensatz tai12, " +
u'12 Flüge, 12 Gates, keine Kosten')
tikz_save('mms_given.tikz',
figureheight='\\figureheight',
figurewidth='\\figurewidth')
#plt.show()
return alldata, n_deleted, v_deleted
def example_instance_big():
nP = 3
jobs = [5, 7, 3, 5, 9, 6, 3, 1, 3, 1]
diff = [1, 2, 2, 5]
base_sampler = DWaveSampler()
nodelist = base_sampler.nodelist
edgelist = base_sampler.edgelist
Q, num_entries, max_value = create.createQUBO(jobs, nP, diff)
embedding = {
0: [
622, 614, 606, 598, 726, 594, 722, 978, 630, 638, 850, 593, 465,
337, 602, 1106, 590
],
1: [
1090, 834, 1110, 578, 1102, 1094, 962, 1118, 706, 1134, 1126, 1142,
1150, 710, 836, 718, 1112, 702, 583
],
2: [894, 830, 838, 846, 854, 886, 878, 870, 862, 851, 723, 595, 979],
3: [362, 106, 490, 234, 618, 749, 1002, 746, 874, 110, 118, 126, 1130],
4: [
342, 334, 326, 318, 382, 374, 569, 953, 366, 358, 350, 313, 441,
697, 825, 1081
],
5: [
315, 325, 341, 333, 317, 365, 357, 349, 443, 571, 699, 381, 827,
955, 373
],
6: [123, 635, 379, 1019, 763, 251, 891, 1147, 507, 1275, 1151],
7: [764, 756, 716, 748, 740, 732, 724, 708, 700],
8: [
591, 599, 607, 345, 985, 601, 729, 857, 615, 473, 639, 1113, 623,
631
],
9: [
511, 455, 704, 448, 471, 463, 960, 576, 479, 487, 496, 495, 503,
320, 624, 832, 1088
],
10: [
709, 1141, 971, 1099, 1109, 843, 717, 844, 715, 701, 1101, 1117,
1125, 1133, 725, 733, 1149, 741
],
11: [761, 1017, 633, 889, 505, 377, 249, 1145, 121, 1273, 127],
12: [893, 829, 885, 877, 837, 845, 853, 861, 869, 880, 752, 1008],
13: [1007, 999, 991, 959, 967, 975, 983, 1015, 1023, 977, 849, 721],
14: [
1127, 961, 1119, 833, 1095, 1103, 1111, 1089, 705, 839, 1135, 831,
847, 1120, 1248, 1254, 1262, 1270, 1278, 855, 863
],
15: [
608, 589, 605, 1098, 597, 613, 480, 352, 586, 842, 714, 970, 224,
96, 101, 109, 117, 125
],
16: [611, 355, 483, 227, 253, 245, 237, 229, 739, 867, 871, 995, 1123],
17: [491, 1003, 107, 363, 619, 747, 246, 238, 235, 875, 254, 1131],
18: [482, 758, 766, 750, 742, 994, 738, 610, 866, 1122, 354],
19: [
988, 835, 1004, 996, 980, 707, 964, 972, 992, 864, 868, 876, 884,
892, 956, 963
],
20: [632, 1144, 888, 504, 1016, 248, 376, 760, 1272, 120],
21: [
982, 958, 966, 974, 950, 946, 818, 690, 990, 998, 1006, 1014, 1022,
692
],
22: [
442, 452, 468, 460, 1082, 444, 570, 698, 826, 954, 476, 484, 458,
492, 500, 508, 314
],
23: [728, 216, 344, 856, 472, 236, 228, 220, 600, 244, 984, 734, 252],
24: [
510, 502, 494, 865, 486, 481, 478, 353, 609, 993, 737, 1121, 470,
462
],
25: [
84, 72, 456, 200, 328, 92, 76, 100, 98, 226, 968, 108, 116, 584,
461, 469, 712, 124, 840, 1096
],
26: [703, 751, 719, 727, 767, 759, 743, 735, 711, 695, 736],
27: [475, 493, 485, 477, 987, 603, 731, 859, 501, 509, 347, 1115],
28:
[1011, 629, 371, 621, 627, 637, 755, 243, 499, 883, 1139, 115, 1267],
29: [
1005, 997, 989, 1013, 965, 976, 973, 981, 957, 1021, 848, 720, 592,
596, 588
],
30:
[329, 343, 335, 585, 383, 359, 351, 457, 367, 375, 713, 841, 969, 338],
31: [
340, 332, 324, 364, 348, 356, 316, 312, 440, 568, 696, 824, 952,
372, 380, 1080
],
32: [
1116, 1114, 1124, 1084, 1092, 1100, 1108, 986, 730, 1132, 858,
1140, 1148, 1097, 1083
],
33: [1018, 762, 250, 634, 122, 506, 378, 890, 1020, 1146, 1274],
34: [757, 626, 1010, 754, 882, 498, 1138, 370, 242, 765, 114, 1266],
35:
[1001, 617, 636, 628, 620, 873, 489, 745, 361, 233, 105, 1129, 612],
36: [1009, 625, 881, 497, 369, 241, 895, 753, 887, 1137, 113, 1265],
37: [
255, 247, 872, 239, 488, 616, 232, 360, 744, 1000, 1128, 240, 112,
231
]
}
sampler = dimod.StructureComposite(base_sampler, nodelist, edgelist)
new_sampler = FixedEmbeddingComposite(sampler, embedding)
solutions = {
'energy': [],
"valid": [],
'max_time': [],
'chain_break_fraction': []
}
for k in range(10):
s = solve.solveQUBO(Q, jobs, nP, new_sampler)
solutions['energy'].append(s['energy'])
solutions['valid'].append(s['valid'])
solutions['max_time'].append(s['max_time'])
solutions['chain_break_fraction'].append(s['chain_break_fraction'])
solve.evaluate(solutions)
print(solutions)
print("Durchschnittliche Anteil gebrochender Ketten: ",
sum(solutions['chain_break_fraction']) / 100)
qubo = {(i, i): 0.0
for i in range(len(qubomatrix))
} # necessary to keep the order of the sample columns consistent
for index, value in np.ndenumerate(qubomatrix):
if value != 0:
qubo[index] = value
print('Converted matrix into QUBO for D-Wave:\n', qubo, '\n')
# create a pegasus graph and a simulation
graph = dnx.pegasus_graph(
16
) # stands for a pegasus P6 graph, see https://www.dwavesys.com/sites/default/files/mwj_dwave_qubits2018.pdf
simulator = neal.SimulatedAnnealingSampler(
) # this makes it use a classical simulated annealing sampler
composite = dimod.StructureComposite(
simulator, graph.nodes,
graph.edges) # necessary to make it use the pegasus graph
sampler = LazyFixedEmbeddingComposite(composite) # similar as before
response = sampler.sample_qubo(qubo, num_reads=10000, chain_strength=40)
print('Response:\n', response, '\n')
print('Embedding:\n', sampler.embedding, '\n')
print('Nodelist:\n', sampler.nodelist, '\n')
print('Edgelist:\n', sampler.edgelist, '\n')
print('Adjacency:\n', sampler.adjacency, '\n')
# save results in results.txt
with open('results.txt', 'w') as file:
file.write('energy\tnum_occurrences\tsample\n')
for sample, energy, num_occurrences, cbf in response.data():
newsample = np.array([
# working with different topologies import neal import dimod import dwave_networkx as dnx import networkx as nx import dwave.embedding from dwave.system import DWaveSampler, EmbeddingComposite # Creating a Pegasus Sampler P6 = dnx.pegasus_graph(6) classical_sampler = neal.SimulatedAnnealingSampler() sampler = dimod.StructureComposite(classical_sampler, P6.nodes, P6.edges) # working with embeddings num_variables = 40 embedding = dwave.embedding.chimera.find_clique_embedding(num_variables, 16) toPrint1 = max(len(chain) for chain in embedding.values()) print(toPrint1) num_variables = 40 embedding = dwave.embedding.pegasus.find_clique_embedding(num_variables, 6) toPrint2 = max(len(chain) for chain in embedding.values()) print(toPrint2)
def test_intermediate_composites(self):
child = dimod.StructureComposite(dimod.NullSampler(), [0, 1], [(0, 1)])
intermediate = dimod.TrackingComposite(child)
sampler = EmbeddingComposite(intermediate)
self.assertEqual(sampler.target_structure.nodelist, [0, 1])
def find_loop(unit_size=8, i0=0, j0=0, use_qpu=True):
'''
dwaveの各セルを頂点として、頂点をつなげるループを
見つけようとするモデルを実行する。
結果は、エネルギー最小値にならないため
期待した結果にはならない。
今後のマシンの精度向上に期待。
または、どうにかして頑健なグラフ構造にすべきか。
※注意：マシンは個体ごとにところどころ壊れているので、使用するセルを以下の引数で指定する。
Args:
unit_size:
セルの正方形の辺の長さ
i0:
セルの縦方向のオフセット
j0:
セルの横方向のオフセット
use_qpu:
qpuを使うか指定する
'''
qpu_unit_size = 16
C16 = dnx.chimera_graph(qpu_unit_size)
h, J = V(unit_size=unit_size, i0=i0, j0=j0)
#print(h)
#print(J)
if use_qpu:
sampler = DWaveSampler()
samples = sampler.sample_ising(h, J, num_reads=10, annealing_time=20)
if 0:
count = 0
for s,e,o in samples.data(['sample', 'energy', 'num_occurrences']):
print(e, o)
print_chimera_simple(s)
samples = sampler.sample_ising(h, J,
num_reads=100,
anneal_schedule = [[0.0,1.0],[20.0,0.0],[30.0,0.3],[35.0,0.3],[40.0,1.0]],
reinitialize_state=False,
initial_state=s)
count += 1
if count >= 1 :
break
else:
classical_sampler = neal.SimulatedAnnealingSampler()
sampler = dimod.StructureComposite(classical_sampler, C16.nodes, C16.edges)
samples = sampler.sample_ising(h, J, num_reads=10)
count = 0
for s,e,o in samples.data(['sample', 'energy', 'num_occurrences']):
print(e, o)
print_chimera_simple(s)
count += 1
if count >= 10 :
break
