Known Issues in Gurobi 13.0#

Known Issues in Gurobi 13.0.3#

WLS License Check Failure when Running Parallel Environments

If using a WLS license, running optimizations from multiple environments in parallel may fail with messages such as:

GRBException 10009: No Gurobi license, or license not started or set

Workarounds:

  • Use Gurobi 13.0.2 (recommended).

  • Ensure that each call to optimize() starts at least one second after the previous one.

Known Issues in Gurobi 13.0.2#

Unknown Termination Status in Presence of Dense Columns

Fixed in version 13.0.3 and later

When a model with many dense columns is solved, Gurobi might terminate with:

Barrier performed 0 iterations in X.XX seconds (Y.YY work units)
Unknown termination status

Known Issues in Gurobi 13.0.1#

None

Known Issues in Gurobi 13.0.0#

Incorrect Names and Dictionary Keys Generated by addConstrs

Fixed in version 13.0.1 and later

With Python 3.13, Model.addConstrs incorrectly determines the indices to be used for generated constraint names when iterating over a predefined list of tuples.

keys = [(0, 1), (1, 0), (1, 1)]
model.addConstrs(
    (<expression> for i, j in keys),
    name="C",
)
# Expected names: C[0,1] C[1,0] C[1,1]
#   Actual names: C[]    C[]    C[]

If you encounter any unexpected behaviour from addConstrs we recommend adjusting your code to use Model.addConstr instead, the equivalent code being:

keys = [(0, 1), (1, 0), (1, 1)]
for i, j in keys:
    model.addConstr(
        <expression>,
        name=f"C[{i},{j}]",
    )

Tuning Multi-Objective Models is Restricted

When tuning multi-objective models, setting parameters for each objective individually by using MultiObjSettings or Multi-Objective Environments will prevent the tuner from doing any change to the parameters for the corresponding multiobjective pass.

Instead, the tuner should be able to change other parameters for this pass, as explained in Multi-Objective Models.