def solve(
self,
objectives: List[ObjectiveTerm] = [],
constraints: List[ConstraintTerm] = [],
num_reads=10,
sweeps=1000,
beta_range=(1, 50),
):
pyqubo_model = self.parse_to_pyqubo_model(objectives, constraints)
feed_dict = {}
feed_dict.update({o["label"]: o["weight"] for o in objectives})
feed_dict.update({c["label"]: c["weight"] for c in constraints})
if self.vartype == "SPIN":
linear, quad, offset = pyqubo_model.to_ising(feed_dict=feed_dict)
solution = solve_ising(linear,
quad,
num_reads=num_reads,
sweeps=sweeps,
beta_range=beta_range)
else:
model, offset = pyqubo_model.to_qubo(feed_dict=feed_dict)
solution = solve_qubo(model,
num_reads=num_reads,
sweeps=sweeps,
beta_range=beta_range)
return pyqubo_model.decode_solution(solution,
vartype=self.vartype,
feed_dict=feed_dict)
def get_qubo(num_nodes, num_colors, edges):
x = Array.create('x', (num_nodes, num_colors), "BINARY")
adjacent_const = 0.0
for (i, j) in edges:
for k in range(num_colors):
adjacent_const += Constraint(x[i, k] * x[j, k],
label="adjacent({},{})".format(i, j))
onecolor_const = 0.0
for i in range(num_nodes):
onecolor_const += Constraint(
(Sum(0, num_colors, lambda j: x[i, j]) - 1)**2,
label="onecolor{}".format(i))
# combine the two components
alpha = Placeholder("alpha")
H = alpha * onecolor_const + adjacent_const
model = H.compile()
alpha = 1.0
# search for an alpha for which each node has only one color
print("Broken constraints")
while True:
qubo, offset = model.to_qubo(feed_dict={"alpha": alpha})
solution = solve_qubo(qubo)
decoded_solution, broken_constraints, energy = model.decode_solution(
solution, vartype="BINARY", feed_dict={"alpha": alpha})
print(len(broken_constraints))
if not broken_constraints:
break
for (key, value) in broken_constraints.items():
if 'o' == key[0]: # onecolor
alpha += 0.1
break
if 'a' == key[0]: # adjacent
break
return qubo, decoded_solution
def solve(self,model,shape=None,feed_dict=None):
if feed_dict is None:
feed_dict = {'PTMO': 2,'PDMO': 2,'PMC': 1}
qubo, offset = model.to_qubo(feed_dict=feed_dict)
sol = solve_qubo(qubo)
if shape is None:
np_X = self.decode(sol)
else:
np_X = self.decode_(sol,shape)
return np_X
def test_solve_qubo(self):
model = TestSolver.create_number_partition_model()
qubo, offset = model.to_qubo()
solution = solve_qubo(qubo, num_reads=1, sweeps=10)
self.assertTrue(solution == {
's1': 0,
's2': 0,
's3': 1
} or {
's1': 1,
's2': 1,
's3': 0
})
q = {}
if solution_type == 'D-wave':
q = response.record['sample'][0][:NUM_PACKAGE]
elif solution_type == 'SA':
q = list(decoded_solution['q'].values())
package = []
weight = 0
value = 0
for i in range(NUM_PACKAGE):
if q[i] == 1:
package.append(PACKAGE_NAME[i])
weight += WEIGHT[i]
value += VALUE[i]
print('knapsack <-', package)
print('weight = ', weight, 'g')
print('value = ', value, 'yen')
if solution_type == 'D-wave':
# PyQUBOのQAで解を探索する
sampler = EmbeddingComposite(DWaveSampler())
response = sampler.sample_qubo(qubo)
# 結果出力
print_result(response)
elif solution_type == 'SA':
# PyQUBOのSAで解を探索する
raw_solution = solve_qubo(qubo)
# 得られた結果をデコードする
decoded_solution, broken, energy = model.decode_solution(raw_solution, vartype="BINARY", feed_dict=feed_dict)
print_result(decoded_solution)
if i not in orderList:
print("""
Solutions doesn't fulfill requirements.
There are cities you visit more than once.
""")
sys.exit()
return orderList
if __name__ == "__main__":
cities = 5
citiesLocation, paths = makeDemoData(citiesSize=cities)
hamiltonian = TSPHamiltonian(citiesSize=cities, pathsWeight=paths)
feedDict = {"lamTime": 250.0, "lamVisit": 250.0}
qubo, offset = hamiltonian.compile().to_qubo(feed_dict=feedDict)
rawSolution = pyqubo.solve_qubo(qubo)
decodeSolution, broken, energu = hamiltonian.compile().decode_solution(
rawSolution, vartype="BINARY", feed_dict=feedDict)
orderList = dict2AnsList(solutionsDict=decodeSolution, citiesSize=cities)
orderList = list(map(lambda x: x + 1, orderList))
orderList.append(0)
plotMap(citiesLocation=citiesLocation, orderList=orderList)
sys.exit()
distance += d_ij * x[k, i] * x[(k+1)%n_city, j]
# Construct hamiltonian
A = Placeholder("A")
H = distance + A * (time_const + city_const)
# Compile model
model = H.compile()
# Generate QUBO
feed_dict = {'A': 4.0}
qubo, offset = model.to_qubo(feed_dict=feed_dict)
# solve QUBO with Simulated Annealing (SA) provided by neal.
sol = solve_qubo(qubo)
# decode solution, constraints that are broken and energy value
solution_sim, broken, energy = model.decode_solution(sol, vartype="BINARY", feed_dict=feed_dict)
plot_city(cities, solution_sim["c"])
plt.savefig(path + '/venv/results/test/test_simulator.png')
plt.close()
###############################################################################################################################################################################################
# solve TSP problem on D-Wave machine
###############################################################################################################################################################################################
from dwave.system.samplers import DWaveSampler
from dwave.system.composites import EmbeddingComposite
def solve(numbers):
qbits = pyqubo.Array.create('x', len(numbers), vartype='BINARY')
H = sum([(2 * q - 1) * n for n, q in zip(numbers, qbits)])**2
model = H.compile()
qubo, offset = model.to_qubo()
return pyqubo.solve_qubo(qubo)
def solve(self, file_name, all_jobs, job_num, all_machines, machine_num,
job_ids, request_times, due_times, process_intervals,
processing_cost):
solved = False
k = 0
milp_model, assign, start_time = _create_milp_model(
job_num, machine_num, job_ids, request_times, due_times,
process_intervals, processing_cost)
""" Write a log file, cannot call in main() function """
output_file = os.getcwd(
) + "/logs/hybrid/2nd/hybrid-jss-" + file_name + ".log"
milp_model.setParam(grb.GRB.Param.LogFile, output_file)
""" Hybrid strategy """
try:
while not solved:
print(
"-----------------------------------------ITERATION:--------------------------------",
k + 1)
# milp_model.update()
# milp_model.tune()
milp_model.optimize(callbackTerminate)
print("start time after MILP: {}".format(start_time))
if k > 0:
print("After getting integer cuts %i:" % (k + 1))
if milp_model.status == grb.GRB.Status.OPTIMAL:
""" Check the feasibility subproblem by MILP or CP model """
model, next_sequence, next_ts, job_pairs_apos, z_apos, U = self._feasi_by_milp(
job_num, machine_num, job_ids, request_times,
due_times, process_intervals, assign, start_time)
qubo, offset = model.to_qubo()
bqm = dimod.BinaryQuadraticModel.from_qubo(qubo, offset)
# sampler = EmbeddingComposite(DWaveSampler({"qpu": True}))
# response = sampler.sample(bqm)
# next_sequence = response.first.sample
next_sequence = pyqubo.solve_qubo(qubo)
print("y result: {}".format(next_sequence))
print("next ts: {}".format(next_ts))
# # do IIS (test conflict contraints)
# print('The model is infeasible; computing IIS')
# milp2_model.computeIIS()
# if milp2_model.IISMinimal:
# print('IIS is minimal\n')
# else:
# print('IIS is not minimal\n')
# print('\nThe following constraint(s) cannot be satisfied:')
# for c in milp2_model.getConstrs():
# if c.IISConstr:
# print('%s' % c.constrName)
# Infeasible schedule
print(z_apos, job_pairs_apos)
next_start_time = milp_model.getAttr("X", start_time)
print(next_start_time)
check_feasible = True
for (i, j) in job_pairs_apos:
if z_apos[j] == z_apos[i] and next_sequence[
'y_(%d,%d)' % (i, j)] == 1:
next_start_time[j] = next_start_time[
i] + process_intervals[i][z_apos[i]]
print("new: {}".format(next_start_time))
# Case: same machine, same start time --> not feasible solution
for (i, j) in job_pairs_apos:
if z_apos[j] == z_apos[i] and next_start_time[
i] == next_start_time[j]:
check_feasible = False
print("check feasible: {}".format(check_feasible))
# feasible_schedule = {}
# for (i,j) in job_pairs_apos:
# feasible_schedule[(i,j)] = 0
# if z_apos[j] == z_apos[i] and next_sequence['y_(%d,%d)' % (i,j)] == 1 and next_sequence['y_(%d,%d)' % (i,j)] + next_sequence['y_(%d,%d)' % (j,i)] == 1:
# print(i,j, process_intervals[i][z_apos[i]])
# condition = next_start_time[j] < next_start_time[i] + process_intervals[i][z_apos[i]] - U*(1 - next_sequence['y_(%d,%d)' % (i,j)])
# print("here: {}".format(condition))
# if condition:
# continue
# else:
# feasible_schedule[(i,j)] = 1
#
# print(feasible_schedule)
# print(next_start_time)
# if any(value==1 for key,value in feasible_schedule.items()):
# check_feasible = 1
# print(check_feasible)
# print(next_start_time)
if not check_feasible:
# status2_str = StatusDict[milp2_model.status]
# print("Feasibility was stopped with status %s" %status2_str)
""" Infeasible """
k += 1
""" Adding integer cuts """
# 1. Creation of set of elements x[i,m] == 1
set_xim = [
i for i in range(job_num)
for m in range(machine_num)
if assign[(i, m)].x == 1
]
print(set_xim)
# 2. Add constraint to MILP model
milp_model.addConstr(lhs=grb.quicksum([
assign[(i, m)] for i in set_xim
for m in range(machine_num)
if assign[(i, m)].x == 1
]),
sense=grb.GRB.LESS_EQUAL,
rhs=len(set_xim) - 1,
name="integer cut")
# # do IIS (test conflict contraints)
# print('The model is infeasible; computing IIS')
# milp_model.computeIIS()
# if milp_model.IISMinimal:
# print('IIS is minimal\n')
# else:
# print('IIS is not minimal\n')
# print('\nThe following constraint(s) cannot be satisfied:')
# for c in milp_model.getConstrs():
# if c.IISConstr:
# print('%s' % c.constrName)
else:
""" Feasible """
solved = True
print("Optimal Schedule Cost: %i" %
(milp_model.objVal))
milp_model.printStats()
self._formulate_schedules(all_machines, job_ids,
request_times, due_times,
process_intervals, assign,
next_start_time)
else:
status_str = StatusDict[milp_model.status]
print("Optimization was stopped with status %s" %
status_str)
milp_model.params.DualReductions = 0
milp_model._vars = assign, start_time
milp_model.Params.lazyConstraints = 1
solved = True
except grb.GurobiError as e:
print("Error code " + str(e.errno) + ": " + str(e))
return solved
