Skip to content

Fix error message for LDL factorization failures #2956

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

Merged
odow merged 2 commits into from
Mar 1, 2026
+19 −9
Merged

Fix error message for LDL factorization failures #2956

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e9b6db5
Fix error message for LDL factorization failures
odow Feb 27, 2026
b6487a2
Update
odow Mar 1, 2026
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
6 changes: 5 additions & 1 deletion ext/MathOptInterfaceLDLFactorizationsExt.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,11 @@ function MOI.Bridges.Constraint.compute_sparse_sqrt_fallback(
) where {F<:MOI.ScalarQuadraticFunction,S<:MOI.AbstractSet}
n = LinearAlgebra.checksquare(Q)
factor = LDLFactorizations.ldl(Q)
if minimum(factor.D) < 0
# Ideally we should use `LDLFactorizations.factorized(factor)` here, but it
# has some false negatives. Instead we check that the factorization appeared
# to work. This is a heuristic. There might be other cases where check is
# insufficient.
if minimum(factor.D) < 0 || any(issubnormal, factor.D)
msg = """
Unable to transform a quadratic constraint into a SecondOrderCone
constraint because the quadratic constraint is not convex.
Expand Down
22 changes: 14 additions & 8 deletions test/Bridges/Constraint/test_QuadtoSOCBridge.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -363,11 +363,11 @@ function test_compute_sparse_sqrt_edge_cases()
[1.0 0.0; 0.0 2.0],
# Cholesky works, with pivoting
[1.0 0.0 1.0; 0.0 1.0 1.0; 1.0 1.0 3.0],
# Cholesky fails due to 0 eigen value
# Cholesky fails due to 0 eigen value. LDL works
[1.0 1.0; 1.0 1.0],
# Cholesky succeeds, even though 0 eigen value
[2.0 2.0; 2.0 2.0],
# Cholesky fails because of 0 column/row
# Cholesky fails because of 0 column/row. LDL works
[2.0 0.0; 0.0 0.0],
]
B = SparseArrays.sparse(A)
Expand All @@ -378,12 +378,18 @@ function test_compute_sparse_sqrt_edge_cases()
end
@test isapprox(A, U' * U; atol = 1e-10)
end
A = [-1.0 0.0; 0.0 1.0]
B = SparseArrays.sparse(A)
@test_throws(
MOI.UnsupportedConstraint{typeof(f),typeof(s)},
MOI.Bridges.Constraint.compute_sparse_sqrt(B, f, s),
)
# Test failures
for A in [
[-1.0 0.0; 0.0 1.0],
# Found from test_quadratic_nonconvex_constraint_basic
[0.0 -1.0; -1.0 0.0],
]
B = SparseArrays.sparse(A)
@test_throws(
MOI.UnsupportedConstraint{typeof(f),typeof(s)},
MOI.Bridges.Constraint.compute_sparse_sqrt(B, f, s),
)
end
return
end

Expand Down
Loading