[P]arallel [Hi]gh-order [Li]brary for [P]DEs  Latest
Parallel High-Order Library for PDEs through hp-adaptive Discontinuous Galerkin methods
Public Member Functions | Private Types | Private Member Functions | Private Attributes | List of all members
PHiLiP::ConstraintFromObjective_SimOpt< Real > Class Template Reference

Creates a constraint from an objective function and a offset value. More...

#include <constraintfromobjective_simopt.hpp>

Inheritance diagram for PHiLiP::ConstraintFromObjective_SimOpt< Real >:
Collaboration diagram for PHiLiP::ConstraintFromObjective_SimOpt< Real >:

Public Member Functions

 ConstraintFromObjective_SimOpt (const ROL::Ptr< ROL::Objective_SimOpt< Real > > &obj, const Real offset=0)
 Constructor.
 
const ROL::Ptr< ROL::Objective_SimOpt< Real > > getObjective (void) const
 Get underlying objective function.
 
void setParameter (const std::vector< Real > &param) override
 Set parameter on constraint.
 
void update (const V &des_var_sim, const V &des_var_ctl, bool flag=true, int iter=-1) override
 Update constraint (underlying objective)'s design varibles.
 
void value (V &c, const V &des_var_sim, const V &des_var_ctl, Real &tol) override
 Returns the constraint value equal to the objective value minus the offset.
 
void applyJacobian_1 (V &jacobian_vector, const V &v, const V &des_var_sim, const V &des_var_ctl, Real &tol) override
 Returns dIdX1.dot(v) where v has the size of X1.
 
void applyJacobian_2 (V &jacobian_vector, const V &v, const V &des_var_sim, const V &des_var_ctl, Real &tol) override
 Returns dIdX2.dot(v) where v has the size of X2.
 
void applyAdjointJacobian_1 (V &ajv, const V &v, const V &des_var_sim, const V &des_var_ctl, Real &tol) override
 Returns v*dIdX1 where v should be a scalar.
 
void applyAdjointJacobian_2 (V &ajv, const V &v, const V &des_var_sim, const V &des_var_ctl, Real &tol) override
 Returns v*dIdX2 where v should be a scalar.
 
void applyAdjointHessian_11 (ROL::Vector< double > &output_vector, const ROL::Vector< double > &dual, const ROL::Vector< double > &input_vector, const ROL::Vector< double > &des_var_sim, const ROL::Vector< double > &des_var_ctl, double &tol)
 Returns dual*dIdX1dX1*input_vector where dual should be a scalar, and input_vector has size X1.
 
void applyAdjointHessian_12 (ROL::Vector< double > &output_vector, const ROL::Vector< double > &dual, const ROL::Vector< double > &input_vector, const ROL::Vector< double > &des_var_sim, const ROL::Vector< double > &des_var_ctl, double &tol)
 Returns dual*dIdX2dX1*input_vector where dual should be a scalar, and input_vector has size X1.
 
void applyAdjointHessian_21 (ROL::Vector< double > &output_vector, const ROL::Vector< double > &dual, const ROL::Vector< double > &input_vector, const ROL::Vector< double > &des_var_sim, const ROL::Vector< double > &des_var_ctl, double &tol)
 Returns dual*dIdX1dX2*input_vector where dual should be a scalar, and input_vector has size X2.
 
void applyAdjointHessian_22 (ROL::Vector< double > &output_vector, const ROL::Vector< double > &dual, const ROL::Vector< double > &input_vector, const ROL::Vector< double > &des_var_sim, const ROL::Vector< double > &des_var_ctl, double &tol)
 Returns dual*dIdX2dX2*input_vector where dual should be a scalar, and input_vector has size X2.
 

Private Types

using V = ROL::Vector< Real >
 Shorthand for ROL::Vector.
 

Private Member Functions

Real getValue (const V &x)
 Get a vector's entry value, where the vector is a SingletonVector.
 
void setValue (V &x, Real val)
 Set a value on the vector, where the vector is a SingletonVector.
 

Private Attributes

const ROL::Ptr< ROL::Objective_SimOpt< Real > > obj_
 Underlying objective function.
 
ROL::Ptr< V > dualVector_1_
 Vector of size n_sim to store dIdX1.
 
ROL::Ptr< V > dualVector_2_
 Vector of size n_ctl to store dIdX2.
 
const Real offset_
 Offset value defining the constraint on the objective function.
 
bool isDual1Initialized_
 Check whether dualVector_1_ has been initialized.
 
bool isDual2Initialized_
 Check whether dualVector_2_ has been initialized.
 

Detailed Description

template<class Real>
class PHiLiP::ConstraintFromObjective_SimOpt< Real >

Creates a constraint from an objective function and a offset value.

Same as ROL::ConstraintFromObjective except we do it for the SimOpt version. This will be especially useful for example when we have a target lift constraint such as CL - 0.375 = 0.

Note that Constraints usually provide a vector of constraints with multiple entries and that we rarely need to access 1 specific entry value. However, in the case of an objective, it only contains a single value, hence the use of the SingletonVector. (Weird naming since it has nothing to do with the typical "Singleton" design pattern).

Definition at line 20 of file constraintfromobjective_simopt.hpp.

The documentation for this class was generated from the following file: