Repository navigation
[BUG] model updates don't persist #2171
Description
from cuopt.linear_programming.data_model import DataModel
from cuopt.linear_programming.solver import Solve
from cuopt.linear_programming.solver_settings import SolverMethod, SolverSettings
dm = DataModel()
# min x + 2y, subject to x+y >= 3, 0 <= x <= 2, y free
dm.set_csr_constraint_matrix([1., 1.], [0, 1], [0, 2])
dm.set_objective_coefficients([1., 2.])
dm.set_variable_lower_bounds([0., -float("inf")])
dm.set_variable_upper_bounds([2., float("inf")])
dm.set_constraint_lower_bounds([3.])
dm.set_constraint_upper_bounds([float("inf")])
settings = SolverSettings()
settings.set_parameter("method", SolverMethod.Barrier)
settings.set_parameter("sequence_solve", True)
settings.set_parameter("log_to_console", False)
for rhs in (None, 6., None):
if rhs is not None:
dm.update_rhs([rhs])
print(Solve(dm, settings).get_primal_objective())yields:
initial Optimal 4.000000016254251
updated Optimal 10.00000009227848
repeated Optimal 4.000000016254251
I expected the third solve's objective to be 10; it's surprising that updates are dropped in subsequent solves.
Activity
yuwenchen95 commentedon Oct 19, 2026
The issue looks like the support of range rows
l <= Ax <=uis not covered by update_rhs at present.