Gurobi

Handling Gurobi licenses through the APPSI interface

In order to obtain performance benefits when re-solving a Pyomo model with Gurobi repeatedly, Pyomo has to keep a reference to a gurobipy model between calls to solve(). Depending on the Gurobi license type, this may “consume” a license as long as any APPSI-Gurobi interface exists (i.e., has not been garbage collected). To release a Gurobi license for other processes, use the release_license() method as shown below. Note that release_license() must be called on every instance for this to actually release the license. However, releasing the license will delete the gurobipy model which will have to be reconstructed from scratch the next time solve() is called, negating any performance benefit of the persistent solver interface.

>>> opt = appsi.solvers.Gurobi()  
>>> results = opt.solve(model)  
>>> opt.release_license()  

Also note that both the available() and solve() methods will construct a gurobipy model, thereby (depending on the type of license) “consuming” a license. The available() method has to do this so that the availability does not change between calls to available() and solve(), leading to unexpected errors.

class pyomo.contrib.appsi.solvers.gurobi.GurobiResults(solver)

Bases: Results

class pyomo.contrib.appsi.solvers.gurobi.Gurobi(only_child_vars=False)

Bases: PersistentBase, PersistentSolver

Interface to Gurobi

enum Availability(value)

Bases: IntEnum

An enumeration.

Member Type:

int

Valid values are as follows:

NotFound = <Availability.NotFound: 0>

BadVersion = <Availability.BadVersion: -1>

BadLicense = <Availability.BadLicense: -2>

FullLicense = <Availability.FullLicense: 1>

LimitedLicense = <Availability.LimitedLicense: 2>

NeedsCompiledExtension = <Availability.NeedsCompiledExtension: -3>

add_block(block)

add_constraints(cons: List[_GeneralConstraintData])

add_params(params: List[_ParamData])

add_sos_constraints(cons: List[_SOSConstraintData])

add_variables(variables: List[_GeneralVarData])

available()

Test if the solver is available on this system.

Nominally, this will return True if the solver interface is valid and can be used to solve problems and False if it cannot.

Note that for licensed solvers there are a number of “levels” of available: depending on the license, the solver may be available with limitations on problem size or runtime (e.g., ‘demo’ vs. ‘community’ vs. ‘full’). In these cases, the solver may return a subclass of enum.IntEnum, with members that resolve to True if the solver is available (possibly with limitations). The Enum may also have multiple members that all resolve to False indicating the reason why the interface is not available (not found, bad license, unsupported version, etc).

Returns:

available – An enum that indicates “how available” the solver is. Note that the enum can be cast to bool, which will be True if the solver is runable at all and False otherwise.

Return type:

Solver.Availability

cbCut(con)

Add a cut within a callback.

Parameters:

con (pyomo.core.base.constraint._GeneralConstraintData) – The cut to add

cbGet(what)

cbGetNodeRel(vars)

Parameters:

vars (Var or iterable of Var) –

cbGetSolution(vars)

Parameters:

vars (iterable of vars) –

cbLazy(con)

Parameters:

con (pyomo.core.base.constraint._GeneralConstraintData) – The lazy constraint to add

cbSetSolution(vars, solution)

cbUseSolution()

property config: GurobiConfig

An object for configuring solve options.

Returns:

An object for configuring pyomo solve options such as the time limit. These options are mostly independent of the solver.

Return type:

SolverConfig

get_duals(cons_to_load=None)

Declare sign convention in docstring here.

Parameters:

cons_to_load (list) – A list of the constraints whose duals should be loaded. If cons_to_load is None, then the duals for all constraints will be loaded.

Returns:

duals – Maps constraints to dual values

Return type:

dict

get_gurobi_param_info(param)

Get information about a gurobi parameter.

Parameters:

param (str) – The gurobi parameter to get info for. See Gurobi documentation for possible options.

Return type:

six-tuple containing the parameter name, type, value, minimum value, maximum value, and default value.

get_linear_constraint_attr(con, attr)

Get the value of an attribute on a gurobi linear constraint.

Parameters:

• con (pyomo.core.base.constraint._GeneralConstraintData) – The pyomo constraint for which the corresponding gurobi constraint attribute should be retrieved.

• attr (str) – The attribute to get. See the Gurobi documentation

get_model_attr(attr)

Get the value of an attribute on the Gurobi model.

Parameters:

attr (str) – The attribute to get. See Gurobi documentation for descriptions of the attributes.

get_primals(vars_to_load=None, solution_number=0)

get_quadratic_constraint_attr(con, attr)

Get the value of an attribute on a gurobi quadratic constraint.

Parameters:

• con (pyomo.core.base.constraint._GeneralConstraintData) – The pyomo constraint for which the corresponding gurobi constraint attribute should be retrieved.

• attr (str) – The attribute to get. See the Gurobi documentation

get_reduced_costs(vars_to_load=None)

Parameters:

vars_to_load (list) – A list of the variables whose reduced cost should be loaded. If vars_to_load is None, then all reduced costs will be loaded.

Returns:

reduced_costs – Maps variable to reduced cost

Return type:

ComponentMap

get_slacks(cons_to_load=None)

Parameters:

cons_to_load (list) – A list of the constraints whose slacks should be loaded. If cons_to_load is None, then the slacks for all constraints will be loaded.

Returns:

slacks – Maps constraints to slack values

Return type:

dict

get_sos_attr(con, attr)

Get the value of an attribute on a gurobi sos constraint.

Parameters:

• con (pyomo.core.base.sos._SOSConstraintData) – The pyomo SOS constraint for which the corresponding gurobi SOS constraint attribute should be retrieved.

• attr (str) – The attribute to get. See the Gurobi documentation

get_var_attr(var, attr)

Get the value of an attribute on a gurobi var.

Parameters:

• var (pyomo.core.base.var._GeneralVarData) – The pyomo var for which the corresponding gurobi var attribute should be retrieved.

• attr (str) – The attribute to get. See gurobi documentation

property gurobi_options

A dictionary mapping solver options to values for those options. These are solver specific.

Returns:

A dictionary mapping solver options to values for those options

Return type:

dict

is_persistent()

Returns:

is_persistent – True if the solver is a persistent solver.

Return type:

bool

load_vars(vars_to_load=None, solution_number=0)

Load the solution of the primal variables into the value attribute of the variables.

Parameters:

vars_to_load (list) – A list of the variables whose solution should be loaded. If vars_to_load is None, then the solution to all primal variables will be loaded.

release_license()

remove_block(block)

remove_constraints(cons: List[_GeneralConstraintData])

remove_params(params: List[_ParamData])

remove_sos_constraints(cons: List[_SOSConstraintData])

remove_variables(variables: List[_GeneralVarData])

reset()

set_callback(func=None)

Specify a callback for gurobi to use.

Parameters:

func (function) –

The function to call. The function should have three arguments. The first will be the pyomo model being solved. The second will be the GurobiPersistent instance. The third will be an enum member of gurobipy.GRB.Callback. This will indicate where in the branch and bound algorithm gurobi is at. For example, suppose we want to solve

\[ \begin{align}\begin{aligned}min 2*x + y\\s.t.\\ y >= (x-2)**2\\ 0 <= x <= 4\\ y >= 0\\ y integer\end{aligned}\end{align} \]

as an MILP using extended cutting planes in callbacks.

>>> from gurobipy import GRB 
>>> import pyomo.environ as pe
>>> from pyomo.core.expr.taylor_series import taylor_series_expansion
>>> from pyomo.contrib import appsi
>>>
>>> m = pe.ConcreteModel()
>>> m.x = pe.Var(bounds=(0, 4))
>>> m.y = pe.Var(within=pe.Integers, bounds=(0, None))
>>> m.obj = pe.Objective(expr=2*m.x + m.y)
>>> m.cons = pe.ConstraintList()  # for the cutting planes
>>>
>>> def _add_cut(xval):
...     # a function to generate the cut
...     m.x.value = xval
...     return m.cons.add(m.y >= taylor_series_expansion((m.x - 2)**2))
...
>>> _c = _add_cut(0)  # start with 2 cuts at the bounds of x
>>> _c = _add_cut(4)  # this is an arbitrary choice
>>>
>>> opt = appsi.solvers.Gurobi()
>>> opt.config.stream_solver = True
>>> opt.set_instance(m) 
>>> opt.gurobi_options['PreCrush'] = 1
>>> opt.gurobi_options['LazyConstraints'] = 1
>>>
>>> def my_callback(cb_m, cb_opt, cb_where):
...     if cb_where == GRB.Callback.MIPSOL:
...         cb_opt.cbGetSolution(vars=[m.x, m.y])
...         if m.y.value < (m.x.value - 2)**2 - 1e-6:
...             cb_opt.cbLazy(_add_cut(m.x.value))
...
>>> opt.set_callback(my_callback)
>>> res = opt.solve(m) 

set_gurobi_param(param, val)

Set a gurobi parameter.

Parameters:

• param (str) – The gurobi parameter to set. Options include any gurobi parameter. Please see the Gurobi documentation for options.

• val (any) – The value to set the parameter to. See Gurobi documentation for possible values.

set_instance(model)

set_linear_constraint_attr(con, attr, val)

Set the value of an attribute on a gurobi linear constraint.

Parameters:

• con (pyomo.core.base.constraint._GeneralConstraintData) – The pyomo constraint for which the corresponding gurobi constraint attribute should be modified.

• attr (str) –

  The attribute to be modified. Options are:

  CBasis DStart Lazy

• val (any) – See gurobi documentation for acceptable values.

set_objective(obj: _GeneralObjectiveData)

set_var_attr(var, attr, val)

Set the value of an attribute on a gurobi variable.

Parameters:

• var (pyomo.core.base.var._GeneralVarData) – The pyomo var for which the corresponding gurobi var attribute should be modified.

• attr (str) –

  The attribute to be modified. Options are:

  Start VarHintVal VarHintPri BranchPriority VBasis PStart

• val (any) – See gurobi documentation for acceptable values.

solve(model, timer: HierarchicalTimer | None = None) → Results

Solve a Pyomo model.

Parameters:

• model (_BlockData) – The Pyomo model to be solved

• timer (HierarchicalTimer) – An option timer for reporting timing

Returns:

results – A results object

Return type:

Results

property symbol_map

update(timer: HierarchicalTimer | None = None)

property update_config

update_params()

update_variables(variables: List[_GeneralVarData])

version()

Returns:

version – A tuple representing the version

Return type:

tuple

write(filename)

Write the model to a file (e.g., and lp file).

Parameters:

filename (str) – Name of the file to which the model should be written.