Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Identify squared linear sums as quadratic with generate_standard_repn #1493

Merged
merged 8 commits into from
Jun 17, 2020
Merged

Identify squared linear sums as quadratic with generate_standard_repn #1493

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
4a56ddf
Generate quadratic expression linear sum squared.
johwiebe Jun 9, 2020
9cbc8f5
Change test for (m.a + m.b)**2.
johwiebe Jun 10, 2020
9db6a31
Add test for (linear sum)**2.
johwiebe Jun 10, 2020
38d5928
Remove comments.
johwiebe Jun 10, 2020
cdd8f97
Make sure ids are ordered.
johwiebe Jun 10, 2020
eb39d4c
ensure order of iteration when building quadratics
carldlaird Jun 12, 2020
d2c560e
reversed the sorting to have a consistent key ordering and added a co…
carldlaird Jun 15, 2020
529766f
Merge pull request #1 from carldlaird/squared_lin_exp
johwiebe Jun 16, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
27 changes: 20 additions & 7 deletions pyomo/repn/standard_repn.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -734,16 +734,29 @@ def _collect_pow(exp, multiplier, idMap, compute_values, verbose, quadratic):
elif compute_values and len(res.linear) == 0:
return Results(constant=multiplier*res.constant**exponent)
#
# If there is one linear term, then we compute the quadratic expression for it.
# If the base is linear, then we compute the quadratic expression for it.
#
elif len(res.linear) == 1:
key, coef = res.linear.popitem()
var = idMap[key]
else:
ans = Results()
if not (res.constant.__class__ in native_numeric_types and res.constant == 0):
has_constant = (res.constant.__class__
not in native_numeric_types
or res.constant != 0)
if has_constant:
ans.constant = multiplier*res.constant*res.constant
ans.linear[key] = 2*multiplier*coef*res.constant
ans.quadratic[key,key] = multiplier*coef*coef

# this is reversed since we want to pop off the end for efficiency
# and the quadratic terms have a convention that the indexing tuple
# of key1, key2 is such that key1 <= key2
keys = sorted(res.linear.keys(), reverse=True)
while len(keys) > 0:
key1 = keys.pop()
coef1 = res.linear[key1]
if has_constant:
ans.linear[key1] = 2*multiplier*coef1*res.constant
ans.quadratic[key1,key1] = multiplier*coef1*coef1
for key2 in keys:
coef2 = res.linear[key2]
ans.quadratic[key1,key2] = 2*multiplier*coef1*coef2
return ans

#
Expand Down
65 changes: 58 additions & 7 deletions pyomo/repn/tests/test_standard.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3141,19 +3141,25 @@ def test_pow5(self):
rep = generate_standard_repn(e, compute_values=False, quadratic=True)
#
self.assertFalse( rep.is_fixed() )
self.assertEqual( rep.polynomial_degree(), None )
self.assertEqual( rep.polynomial_degree(), 2 )
self.assertFalse( rep.is_constant() )
self.assertFalse( rep.is_linear() )
self.assertFalse( rep.is_quadratic() )
self.assertTrue( rep.is_quadratic() )
self.assertTrue( rep.is_nonlinear() )
#
self.assertTrue(len(rep.linear_vars) == 0)
self.assertTrue(len(rep.linear_coefs) == 0)
self.assertTrue(len(rep.quadratic_vars) == 0)
self.assertTrue(len(rep.quadratic_coefs) == 0)
self.assertFalse(rep.nonlinear_expr is None)
self.assertTrue(len(rep.nonlinear_vars) == 2)
baseline = { }
self.assertTrue(len(rep.quadratic_vars) == 3)
self.assertTrue(len(rep.quadratic_coefs) == 3)
self.assertTrue(rep.nonlinear_expr is None)
self.assertTrue(len(rep.nonlinear_vars) == 0)
baseline = { (id(m.a), id(m.a)): 1,
(id(m.b), id(m.b)): 1 }
if id(m.a) < id(m.b):
baseline[id(m.a), id(m.b)] = 2
else:
baseline[id(m.b), id(m.a)] = 2

self.assertEqual(baseline, repn_to_dict(rep))

e = (m.a+3)**2
Expand Down Expand Up @@ -3258,6 +3264,51 @@ def test_pow6(self):
baseline = { None:8 }
self.assertEqual(baseline, repn_to_dict(rep))

def test_pow_of_lin_sum(self):
m = ConcreteModel()
m.x = Var(range(4))
e = sum(x for x in m.x.values())**2

rep = generate_standard_repn(e, compute_values=False, quadratic=False)
#
self.assertFalse( rep.is_fixed() )
self.assertEqual( rep.polynomial_degree(), None )
self.assertFalse( rep.is_constant() )
self.assertFalse( rep.is_linear() )
self.assertFalse( rep.is_quadratic() )
self.assertTrue( rep.is_nonlinear() )
#
self.assertTrue(len(rep.linear_vars) == 0)
self.assertTrue(len(rep.linear_coefs) == 0)
self.assertTrue(len(rep.quadratic_vars) == 0)
self.assertTrue(len(rep.quadratic_coefs) == 0)
self.assertFalse(rep.nonlinear_expr is None)
self.assertTrue(len(rep.nonlinear_vars) == 4)
baseline = { }
self.assertEqual(baseline, repn_to_dict(rep))

rep = generate_standard_repn(e, compute_values=False, quadratic=True)
#
self.assertFalse( rep.is_fixed() )
self.assertEqual( rep.polynomial_degree(), 2 )
self.assertFalse( rep.is_constant() )
self.assertFalse( rep.is_linear() )
self.assertTrue( rep.is_quadratic() )
self.assertTrue( rep.is_nonlinear() )
#
self.assertTrue(len(rep.linear_vars) == 0)
self.assertTrue(len(rep.linear_coefs) == 0)
self.assertTrue(len(rep.quadratic_vars) == 10)
self.assertTrue(len(rep.quadratic_coefs) == 10)
self.assertTrue(rep.nonlinear_expr is None)
self.assertTrue(len(rep.nonlinear_vars) == 0)
baseline = {(id(i), id(j)): 2
for i in m.x.values()
for j in m.x.values()
if id(i) < id(j)}
baseline.update({(id(i), id(i)): 1 for i in m.x.values()})
self.assertEqual(baseline, repn_to_dict(rep))

def test_fixed_exponent(self):
m = ConcreteModel()
m.x = Var()
Expand Down