def test_decompose1(self):
model = MPModelBuilder.build_from_scratch(
[8, 2, 5, 5, 8, 3, 9, 7, 6],
[[2, 3, 4, 1, 0, 0, 0, 0, 0],
[1, 2, 3, 2, 0, 0, 0, 0, 0],
[0, 0, 1, 4, 3, 4, 2, 0, 0],
[0, 0, 2, 1, 1, 2, 5, 0, 0],
[0, 0, 0, 0, 0, 0, 2, 1, 2],
[0, 0, 0, 0, 0, 0, 3, 4, 1]],
['L'] * 6,
[7, 6, 9, 7, 3, 5])
decomposer = finkelstein_qb_decomposer.FinkelsteinQBDecomposer(model)
decomposer.decompose()
u = [set([0, 1]), set([2, 3]), set([4, 5])]
s = [set([]), set([2, 3]), set([6])]
m = [set([0, 1]), set([4, 5]), set([8, 7])]
for t, r in ((u, decomposer._u), (s, decomposer._s), (m, decomposer._m)):
self.assert_equal(len(t), len(r))
for i in range(len(t)):
self.assert_equal(t[i], r[i])
tree = decomposer.get_decomposition_tree()
p0 = tree.node[tree.get_root()].get_model()
self.assert_equal([u'x7', u'x8', u'x9'], sorted(p0.get_columns_names()))
p1 = tree.node[tree.neighbors(tree.get_root())[0]].get_model()
self.assert_equal([u'x3', u'x4', u'x5', u'x6', u'x7'],
sorted(p1.get_columns_names()))
def test_decompose1(self):
model = MPModelBuilder.build_from_scratch(
[8, 2, 5, 5, 8, 3, 9, 7, 6],
[[2, 3, 4, 1, 0, 0, 0, 0, 0],
[1, 2, 3, 2, 0, 0, 0, 0, 0],
[0, 0, 1, 4, 3, 4, 2, 0, 0],
[0, 0, 2, 1, 1, 2, 5, 0, 0],
[0, 0, 0, 0, 0, 0, 2, 1, 2],
[0, 0, 0, 0, 0, 0, 3, 4, 1]],
['L'] * 6,
[7, 6, 9, 7, 3, 5])
decomposer = finkelstein_adv_decomposer.FinkelsteinAdvDecomposer(model)
decomposer.decompose()
layer = [[set([0,1,2,3])], [set([2,3,4,5,6])], [set([6,7,8])]]
self.assert_equal(len(layer),len(decomposer._layers))
for i in range(len(layer)):
self.assert_equal(len(layer[i]),len(decomposer._layers[i]))
for j in range(len(layer[i])):
self.assert_equal(layer[i][j],decomposer._layers[i][j])
tree = decomposer.get_decomposition_tree()
p0=tree.node[tree.get_root()].get_model()
self.assert_equal([u'x1',u'x2',u'x3',u'x4'],sorted(p0.get_columns_names()))
p1=tree.node[tree.neighbors(tree.get_root())[0]].get_model()
self.assert_equal([u'x3',u'x4',u'x5',u'x6',u'x7'],sorted(p1.get_columns_names()))
def test_decompose2(self):
model = MPModelBuilder.build_from_scratch(
[1, 1, 1, 1, 1, 1, 1 ],
[[1, 1, 1, 0, 0, 0, 0],
[0, 1, 0, 1, 1, 0, 0],
[0, 0, 1, 0, 0, 1, 1]],
['L'] * 3,
[2, 2, 2])
decomposer = finkelstein_adv_decomposer.FinkelsteinAdvDecomposer(model)
decomposer.decompose()
tree = decomposer.get_decomposition_tree()
layer = [[set([0,1,2])], [set([1,3,4]), set([2,5,6])]]
self.assert_equal(len(layer),len(decomposer._layers))
for i in range(len(layer)):
self.assert_equal(len(layer[i]),len(decomposer._layers[i]))
for layers1, layers2 in ( (layer[i], decomposer._layers[i]) ,(decomposer._layers[i],layer[i]) ):
for j in range(len(layers1)):
ok=False
for k in range(len(layers2)):
if layers2[k]==layers1[j]:
ok=True
if not ok:
print(layers1[j])
print(layers2)
self.assert_equal(ok,True)
def test_decompose2(self):
model = MPModelBuilder.build_from_scratch(
[2, 3, 1, 5, 4, 6, 1],
[[3., 4., 1., 0., 0., 0., 0.],
[0., 2., 3., 3., 0., 0., 0.],
[0., 2., 0., 0., 3., 0., 0.],
[0., 0., 2., 0., 0., 3., 2.]],
['L'] * 4,
[6, 5, 4, 5])
decomposer = finkelstein_qb_decomposer.FinkelsteinQBDecomposer(model)
decomposer.decompose()
u = [set([0]), set([1, 2, 3])]
s = [set([]), set([1, 2])]
m = [set([0]), set([3, 4, 5, 6])]
for t, r in ((u, decomposer._u), (s, decomposer._s), (m, decomposer._m)):
self.assert_equal(len(t), len(r))
for i in range(len(t)):
self.assert_equal(t[i], r[i])
def test_decompose3(self):
model = MPModelBuilder.build_from_scratch(
[1, 1, 1, 1, 1],
[[1, 1, 0, 0, 0],
[0, 1, 1, 0, 0],
[0, 0, 1, 1, 0],
[0, 0, 0, 1, 1]],
['L'] * 4,
[1, 1, 1,1])
decomposer = finkelstein_qb_decomposer.FinkelsteinQBDecomposer(model)
decomposer.decompose()
u = [set([0]), set([1]), set([2]),set([3])]
s = [set([]), set([1]), set([2]),set([3])]
m = [set([0]), set([]), set([]), set([4])]
for t, r in ((u, decomposer._u), (s, decomposer._s), (m, decomposer._m)):
self.assert_equal(len(t), len(r))
for i in range(len(t)):
self.assert_equal(t[i], r[i])
tree = decomposer.get_decomposition_tree()
def test_build(self):
model = MPModelBuilder.build_from_scratch(
[8, 2, 5, 5, 8, 3, 9, 7, 6],
[[2., 3., 4., 1., 0., 0., 0., 0., 0.],
[1., 2., 3., 2., 0., 0., 0., 0., 0.],
[0., 0., 1., 4., 3., 4., 2., 0., 0.],
[0., 0., 2., 1., 1., 2., 5., 0., 0.],
[0., 0., 0., 0., 0., 0., 2., 1., 2.],
[0., 0., 0., 0., 0., 0., 3., 4., 1.]],
["L"] * 6,
[7, 6, 9, 7, 3, 5])
self.assert_equal(9, model.get_num_columns())
self.assert_equal(6, model.get_num_rows())
self.assert_equal(["c1", "c2", "c3", "c4", "c5", "c6"], model.rows_names)
self.assert_equal([8, 2, 5, 5, 8, 3, 9, 7, 6],
model.get_objective_coefficients())
self.assert_equal(["x1", "x2", "x3", "x4", "x5", "x6", "x7", "x8", "x9"],
model.columns_names)
self.assert_equal([0.0] * 9, model.columns_lower_bounds)
self.assert_equal([1.0] * 9, model.columns_upper_bounds)
def run(self): model = MPModelBuilder.build_from_file(self._args.file) model.pprint()
import timeit
import os
from les.ext.google.operations_research.linear_solver import pywraplp
from les.mp_model import MPModelBuilder
from les.backend_solvers.scip import SCIP
directory = '/home/ira/Desktop/imp/diploma/'
#tests = os.listdir(directory)
tests = ['p200x500_10_4_192890124.mps']
for t in tests:
if t.endswith('.mps'):
scip = SCIP()
model = MPModelBuilder.build_from(directory + t)
print t
#'/home/ira/Documents/problems/i6_200x500.mps')
scip.load_model(model)
params = pywraplp.MPSolverParameters()
params.SetIntegerParam(params.PRESOLVE, params.PRESOLVE_OFF)
#params.set_quiet(True)
start_time = timeit.default_timer()
scip.solve(params)
sol = scip.get_solution()
print "Time:", timeit.default_timer() - start_time
print sol.get_objective_value()
for i in sol.get_variables_names():
print i, sol.get_variable_value_by_name(i)