'''
combine the objective and constraints into a single dictionary, which is really a graph, to send to the qpu.
'''
qubo = {**objective, **constraints}
'''
create a sampler, this is what actually solves the issue
'''
sampler = DWaveSampler(solver={"qpu":True})
'''
create an embedding on the qpu automatically for the problem w/ the sampler.
The embedding composite tries to fit the problem on the qpu, which we get from creating a sampler
'''
response = EmbeddingComposite(sampler).sample_qubo(qubo, chain_strength=5, num_reads=100)
'''
print the response, returns in a custom table
'''
print(response.info["problem_id"])
'''
save as json
'''
with open("test.json", "w") as t_r:
json.dump(response.to_serializable(), t_r)
'''
show the data embedded on the qpu.
from job_shop_scheduler import get_jss_bqm
# Construct a BQM for the jobs
jobs = {
"cupcakes": [("mixer", 2), ("oven", 1)],
"smoothie": [("mixer", 1)],
"lasagna": [("oven", 2)]
}
max_time = 4 # Upperbound on how long the schedule can be; 4 is arbitrary
bqm = get_jss_bqm(jobs, max_time)
print(bqm)
# Submit BQM
# Note: may need to tweak the chain strength and the number of reads
sampler = EmbeddingComposite(DWaveSampler(solver={'qpu': True}))
sampleset = sampler.sample(bqm, chain_strength=2, num_reads=1000)
# Grab solution
solution = sampleset.first.sample
# Visualize solution
# Note0: we are making the solution simpler to interpret by restructuring it
# into the following format:
# task_times = {"job": [start_time_for_task0, start_time_for_task1, ..],
# "other_job": [start_time_for_task0, ..]
# ..}
#
# Note1: each node in our BQM is labelled as "<job>_<task_index>,<time>".
# For example, the node "cupcakes_1,2" refers to job 'cupcakes', its 1st task
# (where we are using zero-indexing, so task '("oven", 1)'), starting at time
def __init__(self,
objective_function=None,
dwave_sampler=None,
dwave_sampler_kwargs=None,
num_activation_vectors=None,
activation_vec_hamming_dist=1,
max_hd=None,
parse_samples=True,
experiment_type=None,
num_reads=None,
num_iters=None,
network_type='minimum'):
super().__init__(objective_function=objective_function)
# Initialize switch network:
# The default behavior here is to choose the smaller of either permutation or
# sorting networks for the given input size.
self.n_obj = self.objective_function.n
if network_type == 'sorting':
self.network = SortingNetwork(self.n_obj)
elif network_type == 'permutation':
self.network = PermutationNetwork(self.n_obj)
elif network_type == 'minimum':
s = SortingNetwork(self.n_obj)
p = PermutationNetwork(self.n_obj)
if s.depth <= p.depth:
self.network = s
else:
self.network = p
else:
raise TypeError('Network type {} not recognized'.format(str(network_type)))
self.n_qubo = self.network.depth
self.dwave_solver = None
self.sampler_kwargs = None
self.qpu = False
# Initialize dwave sampler:
if dwave_sampler == 'QPU':
self.dwave_solver = EmbeddingComposite(DWaveSampler())
self.qpu = True
if dwave_sampler_kwargs:
self.sampler_kwargs = dwave_sampler_kwargs
else:
self.sampler_kwargs = dict()
elif dwave_sampler == 'SA':
self.dwave_solver = SimulatedAnnealingSampler()
if num_reads:
self.sampler_kwargs = {
'num_reads': num_reads
}
else:
self.sampler_kwargs = {
'num_reads': 25
}
elif dwave_sampler == 'Tabu':
self.dwave_solver = TabuSampler()
if num_reads:
self.sampler_kwargs = {
'num_reads': num_reads
}
else:
self.sampler_kwargs = {
'num_reads': 250
}
self.stopwatch = 0
# Initialize type of experiment
# When running a timed experiment there is a high number of iterations and a 30 sec wall clock
# When running a iteration experiment there is a iteration limit of 30 and no wall clock
if experiment_type == 'time_lim':
self.n_iters = 1000
self.time_limit = 30
if experiment_type == 'iter_lim' and num_iters:
self.n_iters = num_iters
self.time_limit = False
else:
self.n_iters = 50
self.time_limit = False
if max_hd:
self.max_hd = max_hd
else:
self.max_hd = 0
if num_activation_vectors:
self.num_activation_vec = num_activation_vectors
else:
self.num_activation_vec = self.n_qubo
self.form_qubo = LQUBO(objective_function=self.objective_function,
switch_network=self.network,
max_hamming_dist=self.max_hd,
num_activation_vectors=self.num_activation_vec,
activation_vec_hamming_dist=activation_vec_hamming_dist)
self.solution = self.objective_function.min_v
if parse_samples:
self.selection = CheckAndSelect
else:
self.selection = Select
import networkx as nx w5 = nx.wheel_graph(5) from dwave.system.samplers import DWaveSampler from dwave.system.composites import EmbeddingComposite sampler = EmbeddingComposite(DWaveSampler()) import dwave_networkx as dnx print(dnx.min_vertex_cover(w5, sampler)) print(dnx.min_vertex_cover(w5, sampler))
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 default(cls):
with open('dwave_credentials.txt') as token_file:
sapi_token = token_file.read()
sampler = EmbeddingComposite(DWaveSampler(token=sapi_token, endpoint=cls.DEFAULT_SAPI_URL))
return cls(sampler)
import matplotlib
matplotlib.use("agg")
import matplotlib.pyplot as plt
# Importo l'analizzatore
import dwave.inspector
# Selezioni il sampler che useremo. Voglio un metodo quantistico,
# quindi uso il DwaweSampler
from dwave.system.samplers import DWaveSampler
# Utilizzo un composite per fare l'embedding
from dwave.system.composites import EmbeddingComposite
# Preparo il sampler usando l'embedding scelto
sampler = EmbeddingComposite(DWaveSampler(solver='Advantage_system1.1'))
# Creo un grafo vuoto
G = nx.Graph()
# Prendo un grafo da un file
grafo = open(
"/workspace/Tesi/Esempi-commentati/Pipelines-Antennas/JOHNSON8-2-4.txt",
"rb")
#grafo = open("/workspace/Tesi/Esempi-commentati/Pipelines-Antennas/chesapeake.txt", "rb")
G = nx.read_edgelist(grafo)
#Creo la funzione per formulare il QUBO
def massimo_set_indipendente_qubo(G, weight=None, lagrange=2.0):
# un QUBO vuoto per un grafo vuoto
from collections import defaultdict
from dwave.system.samplers import DwaveSampler
from dwave.system.composites import EmbeddingComposite
G = nx.Graph()
G.add_edges_from([(0, 4), (0, 5), (1, 2), (1, 6), (2, 4), (3, 7), (5, 6),
(6, 7)])
Q = defaultdict(int)
# Constraint
for i in range(8):
Q[(i, 1)] += -7
for j in range(i + 1, 8):
Q[(i, j)] += 2
# Objective
for i, j in G.edges:
Q[(i, i)] += 1
Q[(j, j)] += 1
Q[(i, j)] += -2
sampler = EmbeddingComposite(DWaveSampler(solver='DW_2000Q_6'))
sampleset = sampler.sample_qubo(Q, num_reads=10)
print(sampleset)
"""
Follow the four scenarios listed in task 1 by changing the variables in this file
qubit_1 qubit_2
O -------- O
h_1, J_val, h_2
"""
qubit_1 = 0
qubit_2 = 1
h_1 = 0
h_2 = 0
J_val = -1
h = {qubit_1: h_1, qubit_2: h_2}
J = {(qubit_1, qubit_2): J_val}
sampler = EmbeddingComposite(DWaveSampler(solver=dict(qpu=True)))
response = sampler.sample_ising(h,
J,
num_reads=1000,
num_spin_reversal_transforms=0)
print(response.aggregate())
dwi.show(response)
Q_not = {('x', 'x'): -1, ('x', 'z'): 2, ('z', 'x'): 0, ('z', 'z'): -1}
from dwave.system.samplers import DWaveSampler
from dwave.system.composites import EmbeddingComposite
sampler = DWaveSampler()
sampler_embedded = EmbeddingComposite(sampler)
print(sampler.adjacency[sampler.nodelist[0]])
from dwave.system.composites import FixedEmbeddingComposite
sampler_embedded = FixedEmbeddingComposite(sampler, {'x': [0], 'z': [4]})
print(sampler_embedded.adjacency)
response = sampler_embedded.sample_qubo(Q_not, num_reads=10)
for datum in response.data(['sample', 'energy', 'num_occurrences']):
print(datum.sample, "Energy: ", datum.energy, "Occurrences: ",
datum.num_occurrences)
s5 = nx.star_graph(4) # create a star graph where node 0 is hub to four other nodes.
# Solving Classically on a CPU
from dimod.reference.samplers import ExactSolver
sampler = ExactSolver() # returns the BQM's value for every possible assignment of variable values
import dwave_networkx as dnx
print(dnx.min_vertex_cover(s5, sampler)) # produce a BQM for our s5 graph and solve it on our selected sampler
print('################################################################################')
# Solving on a D-Wave System
from dwave.system.samplers import DWaveSampler
from dwave.system.composites import EmbeddingComposite # EmbeddingComposite(), maps unstructured problems to the graph structure of the selected sampler, a process known as minor-embedding
sampler = EmbeddingComposite(DWaveSampler()) # endpoint='https://URL_to_my_D-Wave_system/', token='ABC-123456789012345678901234567890', solver='My_D-Wave_Solver'
print(dnx.min_vertex_cover(s5, sampler))
print('################################################################################')
w5 = nx.wheel_graph(5) # creates a new graph
print(dnx.min_vertex_cover(w5, sampler)) # solves on a D-Wave system
print(dnx.min_vertex_cover(w5, sampler))
print('################################################################################')
c5 = nx.circular_ladder_graph(5) # replaces the problem graph
print(dnx.min_vertex_cover(c5, sampler)) # submits twice to the D-Wave system for solution
print(dnx.min_vertex_cover(c5, sampler)) # producing two of the possible valid solutions.
from dwave.system.samplers import DWaveSampler
from dwave.system.composites import EmbeddingComposite
import pandas as pd
import numpy as np
import time
import sympy
from sympy import *
sampler = EmbeddingComposite(
DWaveSampler(endpoint='https://cloud.dwavesys.com/sapi',
token='',
solver='DW_2000Q_2_1'))
orig_stdout = sys.stdout
f = open('mapColoring4ColorsDwaveResults.txt', 'w')
sys.stdout = f
print("\n# MAP COLOURING PROBLEM WITH 4 COLOURS ON D-WAVE #\n")
print("\nSymbolic Computing\n")
#### Symbolic Computing
h = 0.0000005 #small number
n = 4 #four colors
def alpha(n, h):
return (h + h)
def beta(n, alfa, h):
return (((n ^ 3 + n ^ 2 + 1) * alfa) + h)
def solve_with_pbruteforce(jobs,
solution,
qpu=False,
num_reads=2000,
max_time=None,
window_size=5,
chain_strength=2,
times=10):
if max_time is None:
max_time = get_result(jobs, solution) + 3
for iteration_number in range(times):
print(iteration_number)
try:
if qpu:
sampler = EmbeddingComposite(
DWaveSampler(solver={'qpu': True}))
else:
sampler = neal.SimulatedAnnealingSampler()
for i in range(max_time - window_size):
info = find_time_window(jobs, solution, i, i + window_size)
new_jobs, indexes, disable_till, disable_since, disabled_variables = info
if not bool(new_jobs): # if new_jobs dict is empty
continue
try:
bqm = get_jss_bqm(new_jobs,
window_size + 1,
disable_till,
disable_since,
disabled_variables,
stitch_kwargs={'min_classical_gap': 2})
except ImpossibleBQM:
print('*' * 25 + " It's impossible to construct a BQM " +
'*' * 25)
continue
if qpu:
sampleset = sampler.sample(bqm,
chain_strength=chain_strength,
num_reads=num_reads)
else:
sampleset = sampler.sample(bqm, num_reads=num_reads)
solution1 = sampleset.first.sample
selected_nodes = [
k for k, v in solution1.items()
if v == 1 and not k.startswith('aux')
]
# Parse node information
task_times = {k: [-1] * len(v) for k, v in new_jobs.items()}
for node in selected_nodes:
job_name, task_time = node.rsplit("_", 1)
task_index, start_time = map(int, task_time.split(","))
task_times[int(job_name)][task_index] = start_time
# improving original solution
sol_found = deepcopy(solution)
for job, times in task_times.items():
for j in range(len(times)):
sol_found[job][indexes[job]
[j]] = task_times[job][j] + i
if checkValidity(jobs, sol_found):
solution = sol_found
yield solution, i # solution and place in frame
except Exception as e:
# uncomment this if you want to apply some behaviuor when exception occurs
# yield 'ex', 'ex'
print(e)
continue
def test_instantiation_smoketest(self):
sampler = EmbeddingComposite(MockSampler())
dtest.assert_sampler_api(sampler)
def train_model(X_train, y_train, X_test, y_test, lmd):
"""
Train qboost model
:param X_train: train input
:param y_train: train label
:param X_test: test input
:param y_test: test label
:param lmd: lmbda to control regularization term
:return:
"""
NUM_READS = 3000
NUM_WEAK_CLASSIFIERS = 35
# lmd = 0.5
TREE_DEPTH = 3
# define sampler
dwave_sampler = DWaveSampler()
# sa_sampler = micro.dimod.SimulatedAnnealingSampler()
emb_sampler = EmbeddingComposite(dwave_sampler)
N_train = len(X_train)
N_test = len(X_test)
print("\n======================================")
print("Train#: %d, Test: %d" %(N_train, N_test))
print('Num weak classifiers:', NUM_WEAK_CLASSIFIERS)
print('Tree depth:', TREE_DEPTH)
# input: dataset X and labels y (in {+1, -1}
# Preprocessing data
imputer = preprocessing.Imputer()
# scaler = preprocessing.MinMaxScaler()
scaler = preprocessing.StandardScaler()
normalizer = preprocessing.Normalizer()
centerer = preprocessing.KernelCenterer()
# X = imputer.fit_transform(X)
X_train = scaler.fit_transform(X_train)
X_train = normalizer.fit_transform(X_train)
X_train = centerer.fit_transform(X_train)
# X_test = imputer.fit_transform(X_test)
X_test = scaler.fit_transform(X_test)
X_test = normalizer.fit_transform(X_test)
X_test = centerer.fit_transform(X_test)
## Adaboost
print('\nAdaboost')
clf = AdaBoostClassifier(n_estimators=NUM_WEAK_CLASSIFIERS)
# scores = cross_val_score(clf, X, y, cv=5, scoring='accuracy')
print('fitting...')
clf.fit(X_train, y_train)
hypotheses_ada = clf.estimators_
# clf.estimator_weights_ = np.random.uniform(0,1,size=NUM_WEAK_CLASSIFIERS)
print('testing...')
y_train_pred = clf.predict(X_train)
y_test_pred = clf.predict(X_test)
print('accu (train): %5.2f'%(metric(y_train, y_train_pred)))
print('accu (test): %5.2f'%(metric(y_test, y_test_pred)))
# Ensembles of Decision Tree
print('\nDecision tree')
clf2 = WeakClassifiers(n_estimators=NUM_WEAK_CLASSIFIERS, max_depth=TREE_DEPTH)
clf2.fit(X_train, y_train)
y_train_pred2 = clf2.predict(X_train)
y_test_pred2 = clf2.predict(X_test)
print(clf2.estimator_weights)
print('accu (train): %5.2f' % (metric(y_train, y_train_pred2)))
print('accu (test): %5.2f' % (metric(y_test, y_test_pred2)))
# Ensembles of Decision Tree
print('\nQBoost')
DW_PARAMS = {'num_reads': NUM_READS,
'auto_scale': True,
# "answer_mode": "histogram",
'num_spin_reversal_transforms': 10,
# 'annealing_time': 10,
'postprocess': 'optimization',
}
clf3 = QBoostClassifier(n_estimators=NUM_WEAK_CLASSIFIERS, max_depth=TREE_DEPTH)
clf3.fit(X_train, y_train, emb_sampler, lmd=lmd, **DW_PARAMS)
y_train_dw = clf3.predict(X_train)
y_test_dw = clf3.predict(X_test)
print(clf3.estimator_weights)
print('accu (train): %5.2f' % (metric(y_train, y_train_dw)))
print('accu (test): %5.2f' % (metric(y_test, y_test_dw)))
# Ensembles of Decision Tree
print('\nQBoostPlus')
clf4 = QboostPlus([clf, clf2, clf3])
clf4.fit(X_train, y_train, emb_sampler, lmd=lmd, **DW_PARAMS)
y_train4 = clf4.predict(X_train)
y_test4 = clf4.predict(X_test)
print(clf4.estimator_weights)
print('accu (train): %5.2f' % (metric(y_train, y_train4)))
print('accu (test): %5.2f' % (metric(y_test, y_test4)))
print("=============================================")
print("Method \t Adaboost \t DecisionTree \t Qboost \t QboostIt")
print("Train\t %5.2f \t\t %5.2f \t\t\t %5.2f \t\t %5.2f"% (metric(y_train, y_train_pred),
metric(y_train, y_train_pred2),
metric(y_train, y_train_dw),
metric(y_train, y_train4)))
print("Test\t %5.2f \t\t %5.2f \t\t\t %5.2f \t\t %5.2f"% (metric(y_test, y_test_pred),
metric(y_test,y_test_pred2),
metric(y_test, y_test_dw),
metric(y_test, y_test4)))
print("=============================================")
# plt.subplot(211)
# plt.bar(range(len(y_test)), y_test)
# plt.subplot(212)
# plt.bar(range(len(y_test)), y_test_dw)
# plt.show()
return
# Add constraint that each pair of nodes with a shared edge not both select one color
for neighbor in neighbors:
v, u = neighbor
for i in range(colors):
variables = [v + str(i), u + str(i)]
csp.add_constraint(not_both_1, variables)
# Convert the binary constraint satisfaction problem to a binary quadratic model
bqm = dwavebinarycsp.stitch(csp)
#Set up a solver using the local system’s default D-Wave Cloud Client configuration file
#and sample 50 times
sampler = EmbeddingComposite(
DWaveSampler(endpoint='https://cloud.dwavesys.com/sapi',
token='')) #Paste your token here
#sampler = neal.SimulatedAnnealingSampler()
response = sampler.sample(bqm, num_reads=2000) # doctest: +SKIP
# Plot the lowest-energy sample if it meets the constraints
# print(response)
#print(response.samples())
sample = next(response.samples()) # doctest: +SKIP
if not csp.check(sample): # doctest: +SKIP
print("Failed to color map")
else:
plot_map(sample)
scaled_quadratic = {
key: quadratic[key] / scaling_factor
for key in quadratic
}
return scaled_linear, scaled_quadratic
scaled_linear, scaled_quadratic = scale_bias_couplings(linear, quadratic)
bqm = dimod.BinaryQuadraticModel(scaled_linear, scaled_quadratic, 0.0,
dimod.BINARY)
reads = 1000
sol_limit = 64
system = EmbeddingComposite(DWaveSampler(solver='DW_2000Q_2_1'))
sampler = dwave_qbsolv.QBSolv()
Tref = time()
response = sampler.sample(bqm,
num_reads=reads,
solver='tabu',
solver_limit=sol_limit,
verbosity=0)
#response = system.sample(bqm, num_reads=reads)
Tfin = time()
# BEST SOLUTION AVAILABLE RN
best_solution = response.first
solution_data = list(dict(best_solution[0]).values())
# Matrix form of solution
def try_k_coloring(k_colors, graph_name, is_simulated=False):
print("processing ", graph_name, "... ")
num_vertices, num_edges, list_edges = read_graph(graph_name)
vertices = [str(i + 1) for i in range(num_vertices)]
csp = dwavebinarycsp.ConstraintSatisfactionProblem(dwavebinarycsp.BINARY)
one_color_configurations = set()
def not_same_color(v1, v2):
#constraint: not to adyacent nodes share same color
return not (v1 and v2)
for i in range(k_colors):
one_color_configurations.add(
tuple(1 if i == j else 0 for j in range(k_colors)))
#constraint: just one color per vertex
for vertex in vertices:
variables = [vertex + "c" + str(i) for i in range(k_colors)]
csp.add_constraint(one_color_configurations, variables)
for edge in list_edges:
v1, v2 = edge
for i in range(k_colors):
variables = [str(v1) + "c" + str(i), str(v2) + "c" + str(i)]
csp.add_constraint(not_same_color, variables)
def plot_map(self):
G = nx.Graph()
G.add_nodes_from(vertices)
G.add_edges_from(list_edges)
# Translate from binary to integer color representation
color_map = {}
for province in vertices:
for i in range(k_colors):
if sample[province + "c" + str(i)]:
color_map[province] = i
# Plot the sample with color-coded nodes
node_colors = [color_map.get(node) for node in G.nodes()]
nx.draw_circular(G,
with_labels=True,
node_color=node_colors,
node_size=3000,
cmap=plt.cm.rainbow)
plt.show()
bqm = dwavebinarycsp.stitch(csp)
counts = []
idSample = 0
if not is_simulated:
client_qpu = Client.from_config()
client_cpu = Client.from_config(profile='prod')
# Set up a solver using the local system’s default D-Wave Cloud Client configuration file
# and sample 50 times
sampler = EmbeddingComposite(DWaveSampler()) # doctest: +SKIP
start_time = time.time()
response = sampler.sample(bqm, num_reads=1024) # doctest: +SKIP
elapsed_time = time.time() - start_time
for s in response.data():
if not isSampleInSamples(s, counts):
counts.append({
'id': idSample,
'sample': s.sample,
'num_occurrences': s.num_occurrences,
'energy': s.energy
})
idSample += 1
# Plot the lowest-energy sample if it meets the constraints
sample = next(response.samples()) # doctest: +SKIP
if not csp.check(sample): # doctest: +SKIP
print("Failed to color map")
print("execution time: ", elapsed_time)
return counts, response
else:
sampler = neal.SimulatedAnnealingSampler()
start_time = time.time()
response = sampler.sample(bqm, num_reads=1024)
elapsed_time = time.time() - start_time
sample = next(response.samples()) # doctest: +SKIP
for s in response.data():
if not isSampleInSamples(s, counts):
counts.append({
'id': idSample,
'sample': s.sample,
'num_occurrences': s.num_occurrences,
'energy': s.energy
})
idSample += 1
if not csp.check(sample): # doctest: +SKIP
print("Failed to color map")
print("execution time: ", elapsed_time)
return counts, response
print('=============================')
print('Given virtual qubits a0, a1, b0, b1, c0, c1, c2, c3, c4;')
print('list possible solutions where:')
print(' C = A * B, and C = 9 (c3=1, c2=0, c1=0, c0=1)')
print('')
# At the top of this file, set useQpu to True to use a live QPU.
#
# For this tutorial we need a triangular configution, but the physical
# topology of the QPU does not have triangles. There is a technique
# called "embedding" that allows us to map our virtual problem onto a
# physical platform. The concept of embedding is described more
# thoroughly in the embedding tutorials.
if (useQpu):
sampler = DWaveSampler() # live QPU
sampler_embedded = EmbeddingComposite(sampler) # we will need to embed
# See these pages for information on embedding:
# https://docs.dwavesys.com/docs/latest/c_gs_4.html
# https://docs.dwavesys.com/docs/latest/c_handbook_5.html
else:
sampler = SimulatedAnnealingSampler() # simulated quantum annealer
sampler_embedded = sampler # we do not need to embed for a simulation
csp = dwavebinarycsp.ConstraintSatisfactionProblem(dwavebinarycsp.BINARY)
"""
Now we tie together logic gates and operators in the form of
constraints. Listing constraints is an easy way to get a problem onto
the D-Wave, but it is not necessarily an optimal way to do it.
Gates that are allowed include:
### H = -q1 - q2 - q3 - q4 + 2*(q1*q2 + q2*q3 + q3*q4 + q1*q3 + q1*q4 + q2*q4)
linear = {
key: -1
for key in [('q' + str(color), 'q' + str(color)) for color in range(1, 5)]
}
quadratic = {
key: 2
for key in [('q' + str(color1), 'q' + str(color2))
for color1 in range(1, 5) for color2 in range(2, 5)
if color1 < color2]
}
Q = dict(linear)
Q.update(quadratic)
response = EmbeddingComposite(DWaveSamplerInstance).sample_qubo(Q,
num_reads=1000)
for sample in response.data():
print(sample[0], "Energy: ", sample[1], "Occurrences: ", sample[2])
###############################################################################################################################################################################################
### Two regions: A,B; 2 colours: 1,2
### 1: only 1 colour per region can be selected, 2: two regions cannot have the same colour
### H1 = -qA1 - qA2 + 2*qA1*qA2 - qB1 - qB2 + 2*qB1*qB2 #qA1 = region A has colour 1
### H2 = -qA1 - qB1 + 2*qA1*qB1 - qA2 - qB2 + 2*qA2*qB2
### H = H1 + H2 = -2*qA1 -2*qA2 -2*qB1 - 2*qB2 + 2*qA1*qA2 + 2*qB1*qB2 + 2*qA1*qB1 + 2*qA2*qB2
c_linear = {
key: -2
for key in [('q' + region + str(color), 'q' + region + str(color))
for region in ['A', 'B'] for color in range(1, 3)]
}
print('================')
print(' ?????? ')
print(' ?? ?? ')
print(' ?? ?? ?? ')
print(' ?? ?? ?? ')
print(' ?? ?? ')
print(' ?????? ')
print('Flip a bunch of coins and show the distribution.')
print('')
# At the top of this file, set useQpu to True to use a live QPU.
if (useQpu):
sampler = DWaveSampler()
# We need an embedding composite sampler because not all qubits are
# working. A trivial embedding lets us avoid dead qubits.
sampler = EmbeddingComposite(sampler)
else:
sampler = SimulatedAnnealingSampler()
# Initialize a binary quadratic model.
# It will use 2000 qubits. All biases are 0 and all couplings are 0.
bqm = {} # binary quadratic model
distrib = {} # distribution
msg = 'How many coins do you want to flip at the same time?'
try:
coins = raw_input(msg)
except:
try:
coins = input(msg)
except:
hJ_lsq = optimize_lsq(p_true, NODES)
# Define the solver
NUM_READS = 5000
NUM_EPOCHS = 100
lrate = 0.2
DW_PARAMS = {'auto_scale': True, 'num_spin_reversal_transforms': 5}
# define sampler
dwave_sampler = DWaveSampler(solver={'qpu': True})
# Some accounts need to replace this line with the next:
# dwave_sampler = DWaveSampler(token = 'My API Token', solver='Solver Name')
sa_sampler = dimod.SimulatedAnnealingSampler()
emb_sampler = EmbeddingComposite(dwave_sampler)
emb_sampler.parameters
T = 1. / 20
Q_lsq = dict(((key, key), T * value)
for (key, value) in zip(NODES, hJ_lsq[:len(NODES)]))
Q_lsq.update(
dict(((NODES[key[0]], NODES[key[1]]), T * value)
for (key, value) in zip(MORAL_EDGES, hJ_lsq[len(NODES):])))
print('Solving Q_LSQ on QPU...')
response_lsq = emb_sampler.sample_qubo(Q_lsq, num_reads=NUM_READS, **DW_PARAMS)
samples_lsq = np.asarray([[datum.sample[v] for v in NODES]
for datum in response_lsq.data()
import dwave.embedding
from dwave.system.composites import (
EmbeddingComposite,
FixedEmbeddingComposite,
LazyFixedEmbeddingComposite,
LazyEmbeddingComposite,
AutoEmbeddingComposite,
)
from dwave.system.testing import MockDWaveSampler, mock
from dwave.embedding import chain_breaks
from dwave.system.warnings import ChainStrengthWarning
@dimod.testing.load_sampler_bqm_tests(EmbeddingComposite(MockDWaveSampler()))
class TestEmbeddingComposite(unittest.TestCase):
def test_instantiation_smoketest(self):
sampler = EmbeddingComposite(MockDWaveSampler())
dimod.testing.assert_sampler_api(sampler)
def test_sample_ising(self):
sampler = EmbeddingComposite(MockDWaveSampler())
h = {0: -1., 4: 2}
J = {(0, 4): 1.5}
response = sampler.sample_ising(h, J)
# nothing failed and we got at least one response back
colors = len(one_color_configurations)
# Create a binary constraint satisfaction problem
csp = dwavebinarycsp.ConstraintSatisfactionProblem(dwavebinarycsp.BINARY)
# Add constraint that each node (province) select a single color
for province in provinces:
variables = [province+str(i) for i in range(colors)]
csp.add_constraint(one_color_configurations, variables)
# Add constraint that each pair of nodes with a shared edge not both select one color
for neighbor in neighbors:
v, u = neighbor
for i in range(colors):
variables = [v+str(i), u+str(i)]
csp.add_constraint(not_both_1, variables)
# Convert the binary constraint satisfaction problem to a binary quadratic model
bqm = dwavebinarycsp.stitch(csp)
# Set up a solver using the local system’s default D-Wave Cloud Client configuration file
# and sample 5 times
sampler = EmbeddingComposite(DWaveSampler()) # doctest: +SKIP
response = sampler.sample(bqm, num_reads=10) # doctest: +SKIP
# Plot the lowest-energy sample if it meets the constraints
sample = next(response.samples()) # doctest: +SKIP
if not csp.check(sample): # doctest: +SKIP
print("Failed to color map")
else:
plot_map(sample)
def test_instantiation_smoketest(self):
sampler = EmbeddingComposite(MockDWaveSampler())
dimod.testing.assert_sampler_api(sampler)
def solve_with_pbruteforce(jobs,
solution,
qpu=False,
num_reads=2000,
max_time=None,
window_size=5,
chain_strength=2,
num_of_iterations=10,
min_classical_gap=2):
# default, safe value of max_time to give some room for improvement
if max_time is None:
max_time = get_result(jobs, solution) + 3
# main loop, iterates over whole instance
for iteration_number in range(num_of_iterations):
print('-' * 10, f"iteration {iteration_number+1}/{num_of_iterations}",
'-' * 10)
try:
if qpu:
sampler = EmbeddingComposite(DWaveSampler())
else:
sampler = neal.SimulatedAnnealingSampler()
# looping over parts of the instance, solving small sub-instances
# of size window_size
from random import sample
for i in sample(range(max_time - window_size),
len(range(max_time - window_size))):
# cutting out the sub-instance
info = find_time_window(jobs, solution, i, i + window_size)
# new_jobs - tasks present in the sub-instance
# indexes - old (full-instance) indexes of tasks in new_jobs
# disable_till, disable_since and disabled_variables are all
# explained in instance_parser.py
new_jobs, indexes, disable_till, disable_since, disabled_variables = info
if not bool(new_jobs): # if sub-instance is empty
continue
# constructing Binary Quadratic Model
try:
bqm = get_jss_bqm(
new_jobs,
window_size + 1,
disable_till,
disable_since,
disabled_variables,
stitch_kwargs={'min_classical_gap': min_classical_gap})
except ImpossibleBQM:
print('*' * 25 + " It's impossible to construct a BQM " +
'*' * 25)
continue
# reding num_reads responses from the sampler
sampleset = sampler.sample(bqm,
chain_strength=chain_strength,
num_reads=num_reads)
# using the best (lowest energy) sample
solution1 = sampleset.first.sample
# variables that were selected by the sampler
# (apart from the auxiliary variables)
selected_nodes = [
k for k, v in solution1.items()
if v == 1 and not k.startswith('aux')
]
# parsing aquired information
task_times = {k: [-1] * len(v) for k, v in new_jobs.items()}
for node in selected_nodes:
job_name, task_time = node.rsplit("_", 1)
task_index, start_time = map(int, task_time.split(","))
task_times[int(job_name)][task_index] = start_time
# constructing a new solution, improved by the aquired info
# newly scheduled tasks are injected into a full instance
sol_found = deepcopy(solution)
for job, times in task_times.items():
for j in range(len(times)):
sol_found[job][indexes[job]
[j]] = task_times[job][j] + i
# checking if the new, improved solution is valid
if checkValidity(jobs, sol_found):
solution = deepcopy(sol_found)
# solution = sol_found
yield solution, i # solution and current position of window
except Exception as e:
# uncomment this if you want to apply some behaviuor
# in demo.py when exception occurs:
# yield 'ex', 'ex'
print(e)
continue
#CONSTRAINTS
## ------- Run our QUBO on the QPU -------
# Set up QPU parameters
chainstrength = 1500
numruns = 40
gam = 100
Q = {}
for i in range(len(S)):
Q[(i, i)] = (S[i] * S[i] - 30 * S[i]) * gam - C[i]
for i in range(len(S)):
for j in range((i + 1), len(S)):
Q[(i, j)] = (2 * S[i] * S[j] + 2) * gam
# Run the QUBO on the solver from your config file
from dwave.system.samplers import DWaveSampler
from dwave.system.composites import EmbeddingComposite
response = EmbeddingComposite(DWaveSampler()).sample_qubo(
Q, chain_strength=chainstrength, num_reads=numruns)
## ------- Return results to user -------
R = iter(response)
E = iter(response.data())
for line in response:
sample = next(R)
S1 = [S[i] for i in sample if sample[i] > 0]
S0 = [S[i] for i in sample if sample[i] < 1]
print("S1 Sum: ", sum(S1), "\t", S1)
import matplotlib.pyplot as plt
pos = nx.circular_layout(G)
#nx.draw(G,pos,with_labels=True)
nlabels=dict((n,(n,d['weight'])) for n,d in G.nodes(data=True))
#nx.draw(G,pos,labels=nlabels, node_size=2000)
elabels = nx.get_edge_attributes(G,'weight')
#nx.draw_networkx_edge_labels(G,pos,edge_labels=elabels,ax=plt.gca())
#plt.show()
# Parker's direct-to-DWave code
from dwave.system.samplers import DWaveSampler
from dwave.system.composites import EmbeddingComposite
# Set Q for the problem QUBO
linear = {('x0', 'x0'): -1, ('x1', 'x1'): -1, ('x2', 'x2'): -1}
quadratic = {('x0', 'x1'): 2, ('x0', 'x2'): 2, ('x1', 'x2'): 2}
Q = dict(linear)
Q.update(quadratic)
# Try converting G to this format
linear_G = {(n,n):d['weight'] for n,d in G.nodes(data=True)}
quadratic_G = nx.get_edge_attributes(G,'weight') # Note this is the same as elabels
QG = dict(linear_G)
QG.update(quadratic_G)
# Minor-embed and sample 1000 times on a default D-Wave system
#response = EmbeddingComposite(DWaveSampler()).sample_qubo(Q, num_reads=1000)
response = EmbeddingComposite(DWaveSampler()).sample_qubo(QG, num_reads=1000)
for datum in response.data(['sample', 'energy', 'num_occurrences']):
print(datum.sample, "Energy: ", datum.energy, "Occurrences: ", datum.num_occurrences)
def anneal(C_i, C_ij, mu, sigma, l, strength_scale, energy_fraction, ngauges,
max_excited_states):
#Initialising h and J as dictionnaries
h = {}
J = {}
for i in range(len(C_i)):
h_i = -2 * sigma[i] * C_i[i]
for j in range(len(C_ij[0])):
if j > i:
J[(i, j)] = float(2 * C_ij[i][j] * sigma[i] * sigma[j])
h_i += 2 * (sigma[i] * C_ij[i][j] * mu[j])
h[i] = h_i
#applying cutoff
print("Number of J before : " + str(len(J))) #J before cutoff
float_vals = []
for i in J.values():
float_vals.append(i)
cutoff = np.percentile(float_vals, AUGMENT_CUTOFF_PERCENTILE)
to_delete = []
for k, v in J.items():
if v < cutoff:
to_delete.append(k)
for k in to_delete:
del J[k]
print("Number of J after : " + str(len(J))) # J after cutof
new_Q = {}
isingpartial = {}
if FIXING_VARIABLES:
#Optimising heuristically the number of coupling terms
Q, _ = dimod.ising_to_qubo(h, J, offset=0.0)
bqm = dimod.BinaryQuadraticModel.from_qubo(Q, offset=0.0)
simple = dimod.fix_variables(bqm, sampling_mode=False)
if simple == {}:
new_Q = Q
else:
Q_indices = []
for i in Q:
if i in simple.keys():
continue
else:
Q_indices.append(i)
new_Q = {key: Q[key] for key in Q_indices}
print('new length', len(new_Q))
isingpartial = simple
if (not FIXING_VARIABLES) or len(new_Q) > 0:
mapping = []
offset = 0
for i in range(len(C_i)):
if i in isingpartial:
mapping.append(None)
offset += 1
else:
mapping.append(i - offset)
if FIXING_VARIABLES:
new_Q_mapped = {}
for (first, second), val in new_Q.items():
new_Q_mapped[(mapping[first], mapping[second])] = val
h, J, _ = dimod.qubo_to_ising(new_Q_mapped)
#Run gauges
qaresults = []
print("Number of variables to anneal :" + str(len(h)))
for g in range(ngauges):
#Finding embedding
qaresult = []
embedded = False
for attempt in range(5):
a = np.sign(np.random.rand(len(h)) - 0.5)
float_h = []
for i in h.values():
float_h.append(i)
h_gauge = float_h * a
J_gauge = {}
for i in range(len(h)):
for j in range(len(h)):
if (i, j) in J:
J_gauge[(i, j)] = J[(i, j)] * a[i] * a[j]
try:
print("Trying to find embeding")
sampler = EmbeddingComposite(
DWaveSampler(token='secret_token'))
embedded = True
break
except ValueError: # no embedding found
print('no embedding found')
embedded = False
continue
if not embedded:
continue
print("emebeding found")
print("Quantum annealing")
try_again = True
while try_again:
try:
#Annealing, saving energy and sample list
sampleset = sampler.sample_ising(
h_gauge,
J_gauge,
chain_strength=strength_scale,
num_reads=200,
annealing_time=20)
try_again = False
except:
print('runtime or ioerror, trying again')
time.sleep(10)
try_again = True
print("Quantum done")
qaresult.append(sampleset.record[0][0].tolist())
qaresult = np.asarray(qaresult)
qaresult = qaresult * a
qaresults[g * nreads:(g + 1) * nreads] = qaresult
full_strings = np.zeros((len(qaresults), len(C_i)))
full_strings = np.asarray(full_strings)
qaresults = np.asarray(qaresults)
if FIXING_VARIABLES:
j = 0
for i in range(len(C_i)):
if i in isingpartial:
full_strings[:, i] = 2 * isingpartial[i] - 1
else:
full_strings[:, i] = qaresults[:, j]
j += 1
else:
full_strings = qaresults
s = np.asarray(full_strings)
energies = np.zeros(len(qaresults))
s[np.where(s > 1)] = 1.0
s[np.where(s < -1)] = -1.0
bits = len(s[0])
for i in range(bits):
energies += 2 * s[:, i] * (-sigma[i] * C_i[i])
for j in range(bits):
if j > i:
energies += 2 * s[:, i] * s[:, j] * sigma[i] * sigma[
j] * C_ij[i][j]
energies += 2 * s[:, i] * sigma[i] * C_ij[i][j] * mu[j]
unique_energies, unique_indices = np.unique(energies,
return_index=True)
ground_energy = np.amin(unique_energies)
if ground_energy < 0:
threshold_energy = (1 - energy_fraction) * ground_energy
else:
threshold_energy = (1 + energy_fraction) * ground_energy
lowest = np.where(unique_energies < threshold_energy)
unique_indices = unique_indices[lowest]
if len(unique_indices) > max_excited_states:
sorted_indices = np.argsort(
energies[unique_indices])[-max_excited_states:]
unique_indices = unique_indices[sorted_indices]
print("unique indices : ", unique_indices)
print(type(unique_indices[0]))
print(type(full_strings))
final_answers = full_strings[unique_indices]
print('number of selected excited states', len(final_answers))
return final_answers
else:
final_answer = []
print("Evrything resolved by FIXING_VARIABLES")
for i in range(len(C_i)):
if i in isingpartial:
final_answer.append(2 * isingpartial[i] - 1)
final_answer = np.array(final_answer)
return np.array([final_answer])
from dwave.system.samplers import DWaveSampler
from dwave.system.composites import EmbeddingComposite
import setting
linear = {('x0', 'x0'): -1, ('x1', 'x1'): -1, ('x2', 'x2'): -1}
quadratic = {('x0', 'x1'): 2, ('x0', 'x2'): 2, ('x1', 'x2'): 2}
Q = dict(linear)
Q.update(quadratic)
response = EmbeddingComposite(
DWaveSampler(token=setting.tokencode)).sample_qubo(Q, num_reads=1000)
for sample, energy, num_occurrences, chain_break_fraction in list(
response.data()):
print(sample, "Energy: ", energy, "Occurrences: ", num_occurrences)
print("Total_real_time ", response.info["timing"]["total_real_time"], "us")
