bool
RestoRestorationPhase::PerformRestoration()
{
DBG_START_METH("RestoRestorationPhase::PerformRestoration",
dbg_verbosity);
Jnlst().Printf(J_DETAILED, J_MAIN,
"Performing second level restoration phase for current constriant violation %8.2e\n", IpCq().curr_constraint_violation());
DBG_ASSERT(IpCq().curr_constraint_violation()>0.);
// Get a grip on the restoration phase NLP and obtain the pointers
// to the original NLP data
SmartPtr<RestoIpoptNLP> resto_ip_nlp =
static_cast<RestoIpoptNLP*> (&IpNLP());
DBG_ASSERT(dynamic_cast<RestoIpoptNLP*> (&IpNLP()));
SmartPtr<IpoptNLP> orig_ip_nlp =
static_cast<IpoptNLP*> (&resto_ip_nlp->OrigIpNLP());
DBG_ASSERT(dynamic_cast<IpoptNLP*> (&resto_ip_nlp->OrigIpNLP()));
// Get the current point and create a new vector for the result
SmartPtr<const CompoundVector> Ccurr_x =
static_cast<const CompoundVector*> (GetRawPtr(IpData().curr()->x()));
SmartPtr<Vector> new_x = IpData().curr()->x()->MakeNew();
SmartPtr<CompoundVector> Cnew_x =
static_cast<CompoundVector*> (GetRawPtr(new_x));
// The x values remain unchanged
SmartPtr<Vector> x = Cnew_x->GetCompNonConst(0);
x->Copy(*Ccurr_x->GetComp(0));
// ToDo in free mu mode - what to do here?
Number mu = IpData().curr_mu();
// Compute the initial values for the n and p variables for the
// equality constraints
Number rho = resto_ip_nlp->Rho();
SmartPtr<Vector> nc = Cnew_x->GetCompNonConst(1);
SmartPtr<Vector> pc = Cnew_x->GetCompNonConst(2);
SmartPtr<const Vector> cvec = orig_ip_nlp->c(*Ccurr_x->GetComp(0));
SmartPtr<Vector> a = nc->MakeNew();
SmartPtr<Vector> b = nc->MakeNew();
a->Set(mu/(2.*rho));
a->Axpy(-0.5, *cvec);
b->Copy(*cvec);
b->Scal(mu/(2.*rho));
solve_quadratic(*a, *b, *nc);
pc->Copy(*cvec);
pc->Axpy(1., *nc);
DBG_PRINT_VECTOR(2, "nc", *nc);
DBG_PRINT_VECTOR(2, "pc", *pc);
// initial values for the n and p variables for the inequality
// constraints
SmartPtr<Vector> nd = Cnew_x->GetCompNonConst(3);
SmartPtr<Vector> pd = Cnew_x->GetCompNonConst(4);
SmartPtr<Vector> dvec = pd->MakeNew();
dvec->Copy(*orig_ip_nlp->d(*Ccurr_x->GetComp(0)));
dvec->Axpy(-1., *IpData().curr()->s());
a = nd->MakeNew();
b = nd->MakeNew();
a->Set(mu/(2.*rho));
a->Axpy(-0.5, *dvec);
b->Copy(*dvec);
b->Scal(mu/(2.*rho));
solve_quadratic(*a, *b, *nd);
pd->Copy(*dvec);
pd->Axpy(1., *nd);
DBG_PRINT_VECTOR(2, "nd", *nd);
DBG_PRINT_VECTOR(2, "pd", *pd);
// Now set the trial point to the solution of the restoration phase
// s and all multipliers remain unchanged
SmartPtr<IteratesVector> new_trial = IpData().curr()->MakeNewContainer();
new_trial->Set_x(*new_x);
IpData().set_trial(new_trial);
IpData().Append_info_string("R");
return true;
}
bool RestoIterateInitializer::SetInitialIterates()
{
DBG_START_METH("RestoIterateInitializer::SetInitialIterates",
dbg_verbosity);
// Get a grip on the restoration phase NLP and obtain the pointers
// to the original NLP data
SmartPtr<RestoIpoptNLP> resto_ip_nlp =
static_cast<RestoIpoptNLP*> (&IpNLP());
SmartPtr<IpoptNLP> orig_ip_nlp =
static_cast<IpoptNLP*> (&resto_ip_nlp->OrigIpNLP());
SmartPtr<IpoptData> orig_ip_data =
static_cast<IpoptData*> (&resto_ip_nlp->OrigIpData());
SmartPtr<IpoptCalculatedQuantities> orig_ip_cq =
static_cast<IpoptCalculatedQuantities*> (&resto_ip_nlp->OrigIpCq());
// Set the value of the barrier parameter
Number resto_mu;
resto_mu = Max(orig_ip_data->curr_mu(),
orig_ip_cq->curr_c()->Amax(),
orig_ip_cq->curr_d_minus_s()->Amax());
IpData().Set_mu(resto_mu);
Jnlst().Printf(J_DETAILED, J_INITIALIZATION,
"Initial barrier parameter resto_mu = %e\n", resto_mu);
/////////////////////////////////////////////////////////////////////
// Initialize primal varialbes //
/////////////////////////////////////////////////////////////////////
// initialize the data structures in the restoration phase NLP
IpData().InitializeDataStructures(IpNLP(), false, false, false,
false, false);
SmartPtr<Vector> new_x = IpData().curr()->x()->MakeNew();
SmartPtr<CompoundVector> Cnew_x =
static_cast<CompoundVector*> (GetRawPtr(new_x));
// Set the trial x variables from the original NLP
Cnew_x->GetCompNonConst(0)->Copy(*orig_ip_data->curr()->x());
// Compute the initial values for the n and p variables for the
// equality constraints
Number rho = resto_ip_nlp->Rho();
DBG_PRINT((1,"rho = %e\n", rho));
SmartPtr<Vector> nc = Cnew_x->GetCompNonConst(1);
SmartPtr<Vector> pc = Cnew_x->GetCompNonConst(2);
SmartPtr<const Vector> cvec = orig_ip_cq->curr_c();
DBG_PRINT_VECTOR(2, "cvec", *cvec);
SmartPtr<Vector> a = nc->MakeNew();
SmartPtr<Vector> b = nc->MakeNew();
a->Set(resto_mu/(2.*rho));
a->Axpy(-0.5, *cvec);
b->Copy(*cvec);
b->Scal(resto_mu/(2.*rho));
DBG_PRINT_VECTOR(2, "a", *a);
DBG_PRINT_VECTOR(2, "b", *b);
solve_quadratic(*a, *b, *nc);
pc->Copy(*cvec);
pc->Axpy(1., *nc);
DBG_PRINT_VECTOR(2, "nc", *nc);
DBG_PRINT_VECTOR(2, "pc", *pc);
// initial values for the n and p variables for the inequality
// constraints
SmartPtr<Vector> nd = Cnew_x->GetCompNonConst(3);
SmartPtr<Vector> pd = Cnew_x->GetCompNonConst(4);
cvec = orig_ip_cq->curr_d_minus_s();
a = nd->MakeNew();
b = nd->MakeNew();
a->Set(resto_mu/(2.*rho));
a->Axpy(-0.5, *cvec);
b->Copy(*cvec);
b->Scal(resto_mu/(2.*rho));
solve_quadratic(*a, *b, *nd);
pd->Copy(*cvec);
pd->Axpy(1., *nd);
DBG_PRINT_VECTOR(2, "nd", *nd);
DBG_PRINT_VECTOR(2, "pd", *pd);
// Leave the slacks unchanged
SmartPtr<const Vector> new_s = orig_ip_data->curr()->s();
// Now set the primal trial variables
DBG_PRINT_VECTOR(2,"new_s",*new_s);
DBG_PRINT_VECTOR(2,"new_x",*new_x);
SmartPtr<IteratesVector> trial = IpData().curr()->MakeNewContainer();
trial->Set_primal(*new_x, *new_s);
IpData().set_trial(trial);
DBG_PRINT_VECTOR(2, "resto_c", *IpCq().trial_c());
DBG_PRINT_VECTOR(2, "resto_d_minus_s", *IpCq().trial_d_minus_s());
/////////////////////////////////////////////////////////////////////
// Initialize bound multipliers //
/////////////////////////////////////////////////////////////////////
SmartPtr<Vector> new_z_L = IpData().curr()->z_L()->MakeNew();
SmartPtr<CompoundVector> Cnew_z_L =
static_cast<CompoundVector*> (GetRawPtr(new_z_L));
DBG_ASSERT(IsValid(Cnew_z_L));
SmartPtr<Vector> new_z_U = IpData().curr()->z_U()->MakeNew();
SmartPtr<Vector> new_v_L = IpData().curr()->v_L()->MakeNew();
SmartPtr<Vector> new_v_U = IpData().curr()->v_U()->MakeNew();
// multipliers for the original bounds are
SmartPtr<const Vector> orig_z_L = orig_ip_data->curr()->z_L();
SmartPtr<const Vector> orig_z_U = orig_ip_data->curr()->z_U();
SmartPtr<const Vector> orig_v_L = orig_ip_data->curr()->v_L();
SmartPtr<const Vector> orig_v_U = orig_ip_data->curr()->v_U();
// Set the new multipliers to the min of the penalty parameter Rho
// and their current value
SmartPtr<Vector> Cnew_z_L0 = Cnew_z_L->GetCompNonConst(0);
Cnew_z_L0->Set(rho);
Cnew_z_L0->ElementWiseMin(*orig_z_L);
new_z_U->Set(rho);
new_z_U->ElementWiseMin(*orig_z_U);
new_v_L->Set(rho);
new_v_L->ElementWiseMin(*orig_v_L);
new_v_U->Set(rho);
new_v_U->ElementWiseMin(*orig_v_U);
// Set the multipliers for the p and n bounds to the "primal" multipliers
SmartPtr<Vector> Cnew_z_L1 = Cnew_z_L->GetCompNonConst(1);
Cnew_z_L1->Set(resto_mu);
Cnew_z_L1->ElementWiseDivide(*nc);
SmartPtr<Vector> Cnew_z_L2 = Cnew_z_L->GetCompNonConst(2);
Cnew_z_L2->Set(resto_mu);
Cnew_z_L2->ElementWiseDivide(*pc);
SmartPtr<Vector> Cnew_z_L3 = Cnew_z_L->GetCompNonConst(3);
Cnew_z_L3->Set(resto_mu);
Cnew_z_L3->ElementWiseDivide(*nd);
SmartPtr<Vector> Cnew_z_L4 = Cnew_z_L->GetCompNonConst(4);
Cnew_z_L4->Set(resto_mu);
Cnew_z_L4->ElementWiseDivide(*pd);
// Set those initial values to be the trial values in Data
trial = IpData().trial()->MakeNewContainer();
trial->Set_bound_mult(*new_z_L, *new_z_U, *new_v_L, *new_v_U);
IpData().set_trial(trial);
/////////////////////////////////////////////////////////////////////
// Initialize equality constraint multipliers //
/////////////////////////////////////////////////////////////////////
DefaultIterateInitializer::least_square_mults(
Jnlst(), IpNLP(), IpData(), IpCq(),
resto_eq_mult_calculator_, constr_mult_init_max_);
// upgrade the trial to the current point
IpData().AcceptTrialPoint();
DBG_PRINT_VECTOR(2, "y_c", *IpData().curr()->y_c());
DBG_PRINT_VECTOR(2, "y_d", *IpData().curr()->y_d());
DBG_PRINT_VECTOR(2, "z_L", *IpData().curr()->z_L());
DBG_PRINT_VECTOR(2, "z_U", *IpData().curr()->z_U());
DBG_PRINT_VECTOR(2, "v_L", *IpData().curr()->v_L());
DBG_PRINT_VECTOR(2, "v_U", *IpData().curr()->v_U());
return true;
}
void GradientScaling::DetermineScalingParametersImpl(
const SmartPtr<const VectorSpace> x_space,
const SmartPtr<const VectorSpace> p_space,
const SmartPtr<const VectorSpace> c_space,
const SmartPtr<const VectorSpace> d_space,
const SmartPtr<const MatrixSpace> jac_c_space,
const SmartPtr<const MatrixSpace> jac_d_space,
const SmartPtr<const SymMatrixSpace> h_space,
const Matrix& Px_L, const Vector& x_L,
const Matrix& Px_U, const Vector& x_U,
Number& df,
SmartPtr<Vector>& dx,
SmartPtr<Vector>& dc,
SmartPtr<Vector>& dd)
{
DBG_ASSERT(IsValid(nlp_));
SmartPtr<Vector> x = x_space->MakeNew();
SmartPtr<Vector> p = p_space->MakeNew();
if (!nlp_->GetStartingPoint(GetRawPtr(x), true,
GetRawPtr(p), true,
NULL, false,
NULL, false,
NULL, false,
NULL, false)) {
THROW_EXCEPTION(FAILED_INITIALIZATION,
"Error getting initial point from NLP in GradientScaling.\n");
}
//
// Calculate grad_f scaling
//
SmartPtr<Vector> grad_f = x_space->MakeNew();
if (nlp_->Eval_grad_f(*x, *p, *grad_f)) {
double max_grad_f = grad_f->Amax();
df = 1.;
if (scaling_obj_target_gradient_ == 0.) {
if (max_grad_f > scaling_max_gradient_) {
df = scaling_max_gradient_ / max_grad_f;
}
}
else {
if (max_grad_f == 0.) {
Jnlst().Printf(J_WARNING, J_INITIALIZATION,
"Gradient of objective function is zero at starting point. Cannot determine scaling factor based on scaling_obj_target_gradient option.\n");
}
else {
df = scaling_obj_target_gradient_ / max_grad_f;
}
}
df = Max(df, scaling_min_value_);
Jnlst().Printf(J_DETAILED, J_INITIALIZATION,
"Scaling parameter for objective function = %e\n", df);
}
else {
Jnlst().Printf(J_WARNING, J_INITIALIZATION,
"Error evaluating objective gradient at user provided starting point.\n No scaling factor for objective function computed!\n");
df = 1.;
}
//
// No x scaling
//
dx = NULL;
dc = NULL;
if (c_space->Dim()>0) {
//
// Calculate c scaling
//
SmartPtr<Matrix> jac_c = jac_c_space->MakeNew();
if (nlp_->Eval_jac_c(*x, *p, *jac_c)) {
dc = c_space->MakeNew();
const double dbl_min = std::numeric_limits<double>::min();
dc->Set(dbl_min);
jac_c->ComputeRowAMax(*dc, false);
Number arow_max = dc->Amax();
if (scaling_constr_target_gradient_<=0.) {
if (arow_max > scaling_max_gradient_) {
dc->ElementWiseReciprocal();
dc->Scal(scaling_max_gradient_);
SmartPtr<Vector> dummy = dc->MakeNew();
dummy->Set(1.);
dc->ElementWiseMin(*dummy);
}
else {
dc = NULL;
}
}
else {
dc->Set(scaling_constr_target_gradient_/arow_max);
}
if (IsValid(dc) && scaling_min_value_ > 0.) {
SmartPtr<Vector> tmp = dc->MakeNew();
tmp->Set(scaling_min_value_);
dc->ElementWiseMax(*tmp);
}
}
else {
Jnlst().Printf(J_WARNING, J_INITIALIZATION,
"Error evaluating Jacobian of equality constraints at user provided starting point.\n No scaling factors for equality constraints computed!\n");
}
}
dd = NULL;
if (d_space->Dim()>0) {
//
// Calculate d scaling
//
SmartPtr<Matrix> jac_d = jac_d_space->MakeNew();
if (nlp_->Eval_jac_d(*x, *p, *jac_d)) {
dd = d_space->MakeNew();
const double dbl_min = std::numeric_limits<double>::min();
dd->Set(dbl_min);
jac_d->ComputeRowAMax(*dd, false);
Number arow_max = dd->Amax();
if (scaling_constr_target_gradient_<=0.) {
if (arow_max > scaling_max_gradient_) {
dd->ElementWiseReciprocal();
dd->Scal(scaling_max_gradient_);
SmartPtr<Vector> dummy = dd->MakeNew();
dummy->Set(1.);
dd->ElementWiseMin(*dummy);
}
else {
dd = NULL;
}
}
else {
dd->Set(scaling_constr_target_gradient_/arow_max);
}
if (IsValid(dd) && scaling_min_value_ > 0.) {
SmartPtr<Vector> tmp = dd->MakeNew();
tmp->Set(scaling_min_value_);
dd->ElementWiseMax(*tmp);
}
}
else {
Jnlst().Printf(J_WARNING, J_INITIALIZATION,
"Error evaluating Jacobian of inequality constraints at user provided starting point.\n No scaling factors for inequality constraints computed!\n");
}
}
}
void SensAlgorithm::UnScaleIteratesVector(SmartPtr<IteratesVector> *V) {
// unscale the iterates vector
// pretty much a copy from IpOrigIpopt::finalize_solution
SmartPtr<const Vector> unscaled_x;
unscaled_x = IpNLP().NLP_scaling()->unapply_vector_scaling_x((*V)->x());
DBG_ASSERT(IsValid(unscaled_x));
(*V)->Set_x(*unscaled_x);
unscaled_x = NULL ;
SmartPtr<const Matrix> Px_L = IpNLP().Px_L();
SmartPtr<const Matrix> Px_U = IpNLP().Px_U();
SmartPtr<const VectorSpace> x_space = IpNLP().x_space();
SmartPtr<const Vector> y_c = (*V)->y_c();
SmartPtr<const Vector> y_d = (*V)->y_d();
SmartPtr<const Vector> z_L = (*V)->z_L();
SmartPtr<const Vector> z_U = (*V)->z_U();
// unscale y_c
SmartPtr<const Vector> unscaled_yc;
SmartPtr<const Vector> unscaled_yd;
SmartPtr<const Vector> unscaled_z_L;
SmartPtr<const Vector> unscaled_z_U;
Number obj_unscale_factor = IpNLP().NLP_scaling()->unapply_obj_scaling(1.);
if (obj_unscale_factor!=1.) {
SmartPtr<Vector> tmp = IpNLP().NLP_scaling()->apply_vector_scaling_x_LU_NonConst(*Px_L, z_L, *x_space);
tmp->Scal(obj_unscale_factor);
unscaled_z_L = ConstPtr(tmp);
tmp = IpNLP().NLP_scaling()->apply_vector_scaling_x_LU_NonConst(*Px_U, z_U, *x_space);
tmp->Scal(obj_unscale_factor);
unscaled_z_U = ConstPtr(tmp);
tmp = IpNLP().NLP_scaling()->apply_vector_scaling_c_NonConst(y_c);
tmp->Scal(obj_unscale_factor);
unscaled_yc = ConstPtr(tmp);
tmp = IpNLP().NLP_scaling()->apply_vector_scaling_d_NonConst(y_d);
tmp->Scal(obj_unscale_factor);
unscaled_yd = ConstPtr(tmp);
}
else {
unscaled_z_L = IpNLP().NLP_scaling()->apply_vector_scaling_x_LU(*Px_L, z_L, *x_space);
unscaled_z_U = IpNLP().NLP_scaling()->apply_vector_scaling_x_LU(*Px_U, z_U, *x_space);
unscaled_yc = IpNLP().NLP_scaling()->apply_vector_scaling_c(y_c);
unscaled_yd = IpNLP().NLP_scaling()->apply_vector_scaling_d(y_d);
}
(*V)->Set_z_U(*unscaled_z_U);
(*V)->Set_z_L(*unscaled_z_L);
(*V)->Set_y_c(*unscaled_yc);
(*V)->Set_y_d(*unscaled_yd);
}
bool RestoIpoptNLP::InitializeStructures(
SmartPtr<Vector>& x,
bool init_x,
SmartPtr<Vector>& y_c,
bool init_y_c,
SmartPtr<Vector>& y_d,
bool init_y_d,
SmartPtr<Vector>& z_L,
bool init_z_L,
SmartPtr<Vector>& z_U,
bool init_z_U,
SmartPtr<Vector>& v_L,
SmartPtr<Vector>& v_U)
{
DBG_START_METH("RestoIpoptNLP::InitializeStructures", 0); DBG_ASSERT(initialized_);
///////////////////////////////////////////////////////////
// Get the vector/matrix spaces for the original problem //
///////////////////////////////////////////////////////////
SmartPtr<const VectorSpace> orig_x_space;
SmartPtr<const VectorSpace> orig_c_space;
SmartPtr<const VectorSpace> orig_d_space;
SmartPtr<const VectorSpace> orig_x_l_space;
SmartPtr<const MatrixSpace> orig_px_l_space;
SmartPtr<const VectorSpace> orig_x_u_space;
SmartPtr<const MatrixSpace> orig_px_u_space;
SmartPtr<const VectorSpace> orig_d_l_space;
SmartPtr<const MatrixSpace> orig_pd_l_space;
SmartPtr<const VectorSpace> orig_d_u_space;
SmartPtr<const MatrixSpace> orig_pd_u_space;
SmartPtr<const MatrixSpace> orig_jac_c_space;
SmartPtr<const MatrixSpace> orig_jac_d_space;
SmartPtr<const SymMatrixSpace> orig_h_space;
orig_ip_nlp_->GetSpaces(orig_x_space, orig_c_space, orig_d_space, orig_x_l_space, orig_px_l_space, orig_x_u_space,
orig_px_u_space, orig_d_l_space, orig_pd_l_space, orig_d_u_space, orig_pd_u_space, orig_jac_c_space,
orig_jac_d_space, orig_h_space);
// Create the restoration phase problem vector/matrix spaces, based
// on the original spaces (pretty inconvenient with all the
// matrix spaces, isn't it?!?)
DBG_PRINT((1, "Creating the x_space_\n"));
// vector x
Index total_dim = orig_x_space->Dim() + 2 * orig_c_space->Dim() + 2 * orig_d_space->Dim();
x_space_ = new CompoundVectorSpace(5, total_dim);
x_space_->SetCompSpace(0, *orig_x_space);
x_space_->SetCompSpace(1, *orig_c_space); // n_c
x_space_->SetCompSpace(2, *orig_c_space); // p_c
x_space_->SetCompSpace(3, *orig_d_space); // n_d
x_space_->SetCompSpace(4, *orig_d_space); // p_d
DBG_PRINT((1, "Setting the c_space_\n"));
// vector c
//c_space_ = orig_c_space;
c_space_ = new CompoundVectorSpace(1, orig_c_space->Dim());
c_space_->SetCompSpace(0, *orig_c_space);
DBG_PRINT((1, "Setting the d_space_\n"));
// vector d
//d_space_ = orig_d_space;
d_space_ = new CompoundVectorSpace(1, orig_d_space->Dim());
d_space_->SetCompSpace(0, *orig_d_space);
DBG_PRINT((1, "Creating the x_l_space_\n"));
// vector x_L
total_dim = orig_x_l_space->Dim() + 2 * orig_c_space->Dim() + 2 * orig_d_space->Dim();
x_l_space_ = new CompoundVectorSpace(5, total_dim);
x_l_space_->SetCompSpace(0, *orig_x_l_space);
x_l_space_->SetCompSpace(1, *orig_c_space); // n_c >=0
x_l_space_->SetCompSpace(2, *orig_c_space); // p_c >=0
x_l_space_->SetCompSpace(3, *orig_d_space); // n_d >=0
x_l_space_->SetCompSpace(4, *orig_d_space); // p_d >=0
DBG_PRINT((1, "Setting the x_u_space_\n"));
// vector x_U
x_u_space_ = new CompoundVectorSpace(1, orig_x_u_space->Dim());
x_u_space_->SetCompSpace(0, *orig_x_u_space);
DBG_PRINT((1, "Creating the px_l_space_\n"));
// matrix px_l
Index total_rows = orig_x_space->Dim() + 2 * orig_c_space->Dim() + 2 * orig_d_space->Dim();
Index total_cols = orig_x_l_space->Dim() + 2 * orig_c_space->Dim() + 2 * orig_d_space->Dim();
px_l_space_ = new CompoundMatrixSpace(5, 5, total_rows, total_cols);
px_l_space_->SetBlockRows(0, orig_x_space->Dim());
px_l_space_->SetBlockRows(1, orig_c_space->Dim());
px_l_space_->SetBlockRows(2, orig_c_space->Dim());
px_l_space_->SetBlockRows(3, orig_d_space->Dim());
px_l_space_->SetBlockRows(4, orig_d_space->Dim());
px_l_space_->SetBlockCols(0, orig_x_l_space->Dim());
px_l_space_->SetBlockCols(1, orig_c_space->Dim());
px_l_space_->SetBlockCols(2, orig_c_space->Dim());
px_l_space_->SetBlockCols(3, orig_d_space->Dim());
px_l_space_->SetBlockCols(4, orig_d_space->Dim());
px_l_space_->SetCompSpace(0, 0, *orig_px_l_space);
// now setup the identity matrix
// This could be changed to be something like...
// px_l_space_->SetBlockToIdentity(1,1,1.0);
// px_l_space_->SetBlockToIdentity(2,2,other_factor);
// ... etc with some simple changes to the CompoundMatrixSpace
// to allow this (space should auto create the matrices)
//
// for now, we use the new feature and set the true flag for this block
// to say that the matrices should be auto_allocated
SmartPtr<const MatrixSpace> identity_mat_space_nc = new IdentityMatrixSpace(orig_c_space->Dim());
px_l_space_->SetCompSpace(1, 1, *identity_mat_space_nc, true);
px_l_space_->SetCompSpace(2, 2, *identity_mat_space_nc, true);
SmartPtr<const MatrixSpace> identity_mat_space_nd = new IdentityMatrixSpace(orig_d_space->Dim());
px_l_space_->SetCompSpace(3, 3, *identity_mat_space_nd, true);
px_l_space_->SetCompSpace(4, 4, *identity_mat_space_nd, true);
DBG_PRINT((1, "Creating the px_u_space_\n"));
// matrix px_u
total_rows = orig_x_space->Dim() + 2 * orig_c_space->Dim() + 2 * orig_d_space->Dim();
total_cols = orig_x_u_space->Dim();
DBG_PRINT((1, "total_rows = %d, total_cols = %d\n", total_rows, total_cols));
px_u_space_ = new CompoundMatrixSpace(5, 1, total_rows, total_cols);
px_u_space_->SetBlockRows(0, orig_x_space->Dim());
px_u_space_->SetBlockRows(1, orig_c_space->Dim());
px_u_space_->SetBlockRows(2, orig_c_space->Dim());
px_u_space_->SetBlockRows(3, orig_d_space->Dim());
px_u_space_->SetBlockRows(4, orig_d_space->Dim());
px_u_space_->SetBlockCols(0, orig_x_u_space->Dim());
px_u_space_->SetCompSpace(0, 0, *orig_px_u_space);
// other matrices are zero'ed out
// vector d_L
//d_l_space_ = orig_d_l_space;
d_l_space_ = new CompoundVectorSpace(1, orig_d_l_space->Dim());
d_l_space_->SetCompSpace(0, *orig_d_l_space);
// vector d_U
//d_u_space_ = orig_d_u_space;
d_u_space_ = new CompoundVectorSpace(1, orig_d_u_space->Dim());
d_u_space_->SetCompSpace(0, *orig_d_u_space);
// matrix pd_L
//pd_l_space_ = orig_pd_l_space;
pd_l_space_ = new CompoundMatrixSpace(1, 1, orig_pd_l_space->NRows(), orig_pd_l_space->NCols());
pd_l_space_->SetBlockRows(0, orig_pd_l_space->NRows());
pd_l_space_->SetBlockCols(0, orig_pd_l_space->NCols());
pd_l_space_->SetCompSpace(0, 0, *orig_pd_l_space);
// matrix pd_U
//pd_u_space_ = orig_pd_u_space;
pd_u_space_ = new CompoundMatrixSpace(1, 1, orig_pd_u_space->NRows(), orig_pd_u_space->NCols());
pd_u_space_->SetBlockRows(0, orig_pd_u_space->NRows());
pd_u_space_->SetBlockCols(0, orig_pd_u_space->NCols());
pd_u_space_->SetCompSpace(0, 0, *orig_pd_u_space);
DBG_PRINT((1, "Creating the jac_c_space_\n"));
// matrix jac_c
total_rows = orig_c_space->Dim();
total_cols = orig_x_space->Dim() + 2 * orig_c_space->Dim() + 2 * orig_d_space->Dim();
jac_c_space_ = new CompoundMatrixSpace(1, 5, total_rows, total_cols);
jac_c_space_->SetBlockRows(0, orig_c_space->Dim());
jac_c_space_->SetBlockCols(0, orig_x_space->Dim());
jac_c_space_->SetBlockCols(1, orig_c_space->Dim());
jac_c_space_->SetBlockCols(2, orig_c_space->Dim());
jac_c_space_->SetBlockCols(3, orig_d_space->Dim());
jac_c_space_->SetBlockCols(4, orig_d_space->Dim());
jac_c_space_->SetCompSpace(0, 0, *orig_jac_c_space);
// **NOTE: By placing "flat" identity matrices here, we are creating
// potential issues for linalg operations that arise when the original
// NLP has a "compound" c_space. To avoid problems like this,
// we place all unmodified component spaces in trivial (size 1)
// "compound" spaces.
jac_c_space_->SetCompSpace(0, 1, *identity_mat_space_nc, true);
jac_c_space_->SetCompSpace(0, 2, *identity_mat_space_nc, true);
// remaining blocks are zero'ed
DBG_PRINT((1, "Creating the jac_d_space_\n"));
// matrix jac_d
total_rows = orig_d_space->Dim();
total_cols = orig_x_space->Dim() + 2 * orig_c_space->Dim() + 2 * orig_d_space->Dim();
jac_d_space_ = new CompoundMatrixSpace(1, 5, total_rows, total_cols);
jac_d_space_->SetBlockRows(0, orig_d_space->Dim());
jac_d_space_->SetBlockCols(0, orig_x_space->Dim());
jac_d_space_->SetBlockCols(1, orig_c_space->Dim());
jac_d_space_->SetBlockCols(2, orig_c_space->Dim());
jac_d_space_->SetBlockCols(3, orig_d_space->Dim());
jac_d_space_->SetBlockCols(4, orig_d_space->Dim());
jac_d_space_->SetCompSpace(0, 0, *orig_jac_d_space);
DBG_PRINT((1, "orig_jac_d_space = %x\n", GetRawPtr(orig_jac_d_space)))
// Blocks (0,1) and (0,2) are zero'ed out
// **NOTE: By placing "flat" identity matrices here, we are creating
// potential issues for linalg operations that arise when the original
// NLP has a "compound" d_space. To avoid problems like this,
// we place all unmodified component spaces in trivial (size 1)
// "compound" spaces.
jac_d_space_->SetCompSpace(0, 3, *identity_mat_space_nd, true);
jac_d_space_->SetCompSpace(0, 4, *identity_mat_space_nd, true);
DBG_PRINT((1, "Creating the h_space_\n"));
// matrix h
total_dim = orig_x_space->Dim() + 2 * orig_c_space->Dim() + 2 * orig_d_space->Dim();
h_space_ = new CompoundSymMatrixSpace(5, total_dim);
h_space_->SetBlockDim(0, orig_x_space->Dim());
h_space_->SetBlockDim(1, orig_c_space->Dim());
h_space_->SetBlockDim(2, orig_c_space->Dim());
h_space_->SetBlockDim(3, orig_d_space->Dim());
h_space_->SetBlockDim(4, orig_d_space->Dim());
SmartPtr<DiagMatrixSpace> DR_x_space = new DiagMatrixSpace(orig_x_space->Dim());
if( hessian_approximation_ == LIMITED_MEMORY )
{
const LowRankUpdateSymMatrixSpace* LR_h_space = static_cast<const LowRankUpdateSymMatrixSpace*>(GetRawPtr(
orig_h_space));
DBG_ASSERT(LR_h_space);
SmartPtr<LowRankUpdateSymMatrixSpace> new_orig_h_space = new LowRankUpdateSymMatrixSpace(LR_h_space->Dim(),
NULL, orig_x_space, false);
h_space_->SetCompSpace(0, 0, *new_orig_h_space, true);
}
else
{
SmartPtr<SumSymMatrixSpace> sumsym_mat_space = new SumSymMatrixSpace(orig_x_space->Dim(), 2);
sumsym_mat_space->SetTermSpace(0, *orig_h_space);
sumsym_mat_space->SetTermSpace(1, *DR_x_space);
h_space_->SetCompSpace(0, 0, *sumsym_mat_space, true);
// All remaining blocks are zero'ed out
}
///////////////////////////
// Create the bound data //
///////////////////////////
// x_L
x_L_ = x_l_space_->MakeNewCompoundVector();
x_L_->SetComp(0, *orig_ip_nlp_->x_L()); // x >= x_L
x_L_->GetCompNonConst(1)->Set(0.0); // n_c >= 0
x_L_->GetCompNonConst(2)->Set(0.0); // p_c >= 0
x_L_->GetCompNonConst(3)->Set(0.0); // n_d >= 0
x_L_->GetCompNonConst(4)->Set(0.0); // p_d >= 0
DBG_PRINT_VECTOR(2, "resto_x_L", *x_L_);
// x_U
x_U_ = x_u_space_->MakeNewCompoundVector();
x_U_->SetComp(0, *orig_ip_nlp_->x_U());
// d_L
d_L_ = d_l_space_->MakeNewCompoundVector();
d_L_->SetComp(0, *orig_ip_nlp_->d_L());
// d_U
d_U_ = d_u_space_->MakeNewCompoundVector();
d_U_->SetComp(0, *orig_ip_nlp_->d_U());
// Px_L
Px_L_ = px_l_space_->MakeNewCompoundMatrix();
Px_L_->SetComp(0, 0, *orig_ip_nlp_->Px_L());
// Identities are auto-created (true flag passed into SetCompSpace)
// Px_U
Px_U_ = px_u_space_->MakeNewCompoundMatrix();
Px_U_->SetComp(0, 0, *orig_ip_nlp_->Px_U());
// Remaining matrices will be zero'ed out
// Pd_L
//Pd_L_ = orig_ip_nlp_->Pd_L();
Pd_L_ = pd_l_space_->MakeNewCompoundMatrix();
Pd_L_->SetComp(0, 0, *orig_ip_nlp_->Pd_L());
// Pd_U
//Pd_U_ = orig_ip_nlp_->Pd_U();
Pd_U_ = pd_u_space_->MakeNewCompoundMatrix();
Pd_U_->SetComp(0, 0, *orig_ip_nlp_->Pd_U());
// Getting the NLP scaling
SmartPtr<const MatrixSpace> scaled_jac_c_space;
SmartPtr<const MatrixSpace> scaled_jac_d_space;
SmartPtr<const SymMatrixSpace> scaled_h_space;
NLP_scaling()->DetermineScaling(GetRawPtr(x_space_), c_space_, d_space_, GetRawPtr(jac_c_space_),
GetRawPtr(jac_d_space_), GetRawPtr(h_space_), scaled_jac_c_space, scaled_jac_d_space, scaled_h_space, *Px_L_,
*x_L_, *Px_U_, *x_U_);
// For now we assume that no scaling is done inside the NLP_Scaling
DBG_ASSERT(scaled_jac_c_space == jac_c_space_); DBG_ASSERT(scaled_jac_d_space == jac_d_space_); DBG_ASSERT(scaled_h_space == h_space_);
/////////////////////////////////////////////////////////////////////////
// Create and initialize the vectors for the restoration phase problem //
/////////////////////////////////////////////////////////////////////////
// Vector x
SmartPtr<CompoundVector> comp_x = x_space_->MakeNewCompoundVector();
if( init_x )
{
comp_x->GetCompNonConst(0)->Copy(*orig_ip_data_->curr()->x());
comp_x->GetCompNonConst(1)->Set(1.0);
comp_x->GetCompNonConst(2)->Set(1.0);
comp_x->GetCompNonConst(3)->Set(1.0);
comp_x->GetCompNonConst(4)->Set(1.0);
}
x = GetRawPtr(comp_x);
// Vector y_c
y_c = c_space_->MakeNew();
if( init_y_c )
{
y_c->Set(0.0); // ToDo
}
// Vector y_d
y_d = d_space_->MakeNew();
if( init_y_d )
{
y_d->Set(0.0);
}
// Vector z_L
z_L = x_l_space_->MakeNew();
if( init_z_L )
{
z_L->Set(1.0);
}
// Vector z_U
z_U = x_u_space_->MakeNew();
if( init_z_U )
{
z_U->Set(1.0);
}
// Vector v_L
v_L = d_l_space_->MakeNew();
// Vector v_U
v_U = d_u_space_->MakeNew();
// Initialize other data needed by the restoration nlp. x_ref is
// the point to reference to which we based the regularization
// term
x_ref_ = orig_x_space->MakeNew();
x_ref_->Copy(*orig_ip_data_->curr()->x());
dr_x_ = orig_x_space->MakeNew();
dr_x_->Set(1.0);
SmartPtr<Vector> tmp = dr_x_->MakeNew();
tmp->Copy(*x_ref_);
dr_x_->ElementWiseMax(*tmp);
tmp->Scal(-1.);
dr_x_->ElementWiseMax(*tmp);
dr_x_->ElementWiseReciprocal();
DBG_PRINT_VECTOR(2, "dr_x_", *dr_x_);
DR_x_ = DR_x_space->MakeNewDiagMatrix();
DR_x_->SetDiag(*dr_x_);
return true;
}
