def pysp_instance_creation_callback(scenario_name, node_names):
model = create_instance(scenario_name)
#
# SMPS Related Annotations
#
model.constraint_stage = PySP_ConstraintStageAnnotation()
model.constraint_stage.declare(model.c_first_stage, 1)
model.constraint_stage.declare(model.p_first_stage, 1)
model.constraint_stage.declare(model.c_second_stage, 2)
model.constraint_stage.declare(model.r_second_stage, 2)
model.constraint_stage.declare(model.p_second_stage, 2)
# The difficulty with Piecewise is that it hides the
# structure of the underlying constraints (there may be
# more than one). It doesn't seem possible to direct a
# modeler on how to go about tagging specific
# constraints. For this reason, we allow the
# PySP_StochasticRHS and PySP_StochasticMatrix suffixes
# to contain entries for entire blocks, where we
# interpret this as meaning all rhs and constraint
# matrix entries should be treated as stochastic.
model.stoch_rhs = PySP_StochasticRHSAnnotation()
model.stoch_rhs.declare(model.p_second_stage)
model.stoch_rhs.declare(model.r_second_stage)
model.stoch_matrix = PySP_StochasticMatrixAnnotation()
# exercise more of the code by testing this with an
# indexed block and a single block
model.stoch_matrix.declare(model.c_second_stage, variables=[model.r])
model.stoch_matrix.declare(model.p_second_stage[1])
model.stoch_objective = PySP_StochasticObjectiveAnnotation()
return model
def test_deprecated(self):
self.assertIs(StageCostAnnotation, type(PySP_StageCostAnnotation()))
self.assertIs(VariableStageAnnotation,
type(PySP_VariableStageAnnotation()))
self.assertIs(_ConstraintStageAnnotation,
type(PySP_ConstraintStageAnnotation()))
self.assertIs(_ConstraintStageAnnotation,
type(ConstraintStageAnnotation()))
self.assertIs(StochasticDataAnnotation,
type(PySP_StochasticDataAnnotation()))
self.assertIs(StochasticConstraintBoundsAnnotation,
type(PySP_StochasticRHSAnnotation()))
self.assertIs(StochasticConstraintBodyAnnotation,
type(PySP_StochasticMatrixAnnotation()))
self.assertIs(StochasticObjectiveAnnotation,
type(PySP_StochasticObjectiveAnnotation()))
def declare_annotations_rule(model):
#
# Annotate constraint stages
#
model.constraint_stage = PySP_ConstraintStageAnnotation()
# first-stage constraints
model.constraint_stage.declare(model.ConstrainTotalAcreage, 1)
# second-stage constraints
model.constraint_stage.declare(model.EnforceQuotas, 2)
model.constraint_stage.declare(model.EnforceCattleFeedRequirement, 2)
model.constraint_stage.declare(model.LimitAmountSold, 2)
#
# Annotate stochastic constraint matrix coefficients
#
model.stoch_matrix = PySP_StochasticMatrixAnnotation()
# DevotedAcreage[i] has a stochastic coefficient in two constraints
for i in model.CROPS:
model.stoch_matrix.declare(model.EnforceCattleFeedRequirement[i],
variables=[model.DevotedAcreage[i]])
model.stoch_matrix.declare(model.LimitAmountSold[i],
variables=[model.DevotedAcreage[i]])
from pyomo.pysp.annotations import (PySP_ConstraintStageAnnotation,
PySP_StochasticRHSAnnotation)
#
# Import the reference model
#
from baa99_basemodel import model
#
# Annotate the model to enable SMPS conversion
# of the explicit scenario tree defined later
# in this file
#
model.constraint_stage = PySP_ConstraintStageAnnotation()
model.constraint_stage.declare(model.s1, 2)
model.constraint_stage.declare(model.s2, 2)
model.constraint_stage.declare(model.d1, 2)
model.constraint_stage.declare(model.d2, 2)
model.stoch_rhs = PySP_StochasticRHSAnnotation()
model.stoch_rhs.declare(model.d1)
model.stoch_rhs.declare(model.d2)
#
# Define the scenario tree and provide a scenario instance
# creation callback
#
num_scenarios = len(model.d1_rhs_table) * len(model.d2_rhs_table)
scenario_data = dict(
def pysp_instance_creation_callback(scenario_name, node_names):
cnt = d[scenario_name]
model = ConcreteModel()
# first stage
model.x = Var(bounds=(0, 10))
# first stage derived
model.y = Expression(expr=model.x + 1)
model.fx = Var()
# second stage
model.z = Var(bounds=(-10, 10))
# second stage derived
model.q = Expression(expr=model.z**2)
model.fz = Var()
model.r = Var()
# stage costs
model.StageCost = Expression([1, 2])
model.StageCost.add(1, model.fx)
model.StageCost.add(2, -model.fz + model.r - cnt)
model.o = Objective(expr=summation(model.StageCost))
model.ZERO = Expression(expr=0)
if cnt == 0:
cnt = model.ZERO
model.c_first_stage = Constraint(expr=model.x >= 0)
# test our handling of intermediate variables that
# are created by Piecewise but can not necessarily
# be declared on the scenario tree
model.p_first_stage = Piecewise(model.fx,
model.x,
pw_pts=[0., 2., 5., 7., 10.],
pw_constr_type='EQ',
pw_repn='INC',
f_rule=[10., 10., 9., 10., 10.],
force_pw=True)
model.c_second_stage = Constraint(expr=model.x + model.r * cnt >= -100)
model.r_second_stage = Constraint(expr=-cnt <= model.r <= 0)
# exercise more of the code by making this an indexed
# block
model.p_second_stage = Piecewise([1],
model.fz,
model.z,
pw_pts=[-10, -5., 0., 5., 10.],
pw_constr_type='EQ',
pw_repn='INC',
f_rule=[0., 0., -1., 2. + cnt, 1.],
force_pw=True)
#
# SMPS Related Annotations
#
model.constraint_stage = PySP_ConstraintStageAnnotation()
model.constraint_stage.declare(model.c_first_stage, 1)
model.constraint_stage.declare(model.p_first_stage, 1)
model.constraint_stage.declare(model.c_second_stage, 2)
model.constraint_stage.declare(model.r_second_stage, 2)
model.constraint_stage.declare(model.p_second_stage, 2)
# The difficulty with Piecewise is that it hides the
# structure of the underlying constraints (there may be
# more than one). It doesn't seem possible to direct a
# modeler on how to go about tagging specific
# constraints. For this reason, we allow the
# PySP_StochasticRHS and PySP_StochasticMatrix suffixes
# to contain entries for entire blocks, where we
# interpret this as meaning all rhs and constraint
# matrix entries should be treated as stochastic.
model.stoch_rhs = PySP_StochasticRHSAnnotation()
model.stoch_rhs.declare(model.p_second_stage)
model.stoch_rhs.declare(model.r_second_stage)
model.stoch_matrix = PySP_StochasticMatrixAnnotation()
model.stoch_objective = PySP_StochasticObjectiveAnnotation()
# exercise more of the code by testing this with an
# indexed block and a single block
model.stoch_matrix.declare(model.c_second_stage, variables=[model.r])
model.stoch_matrix.declare(model.p_second_stage[1])
return model