MoochoPack_CalcLambdaIndepStd_AddedStep.cpp

#if 0

// @HEADER
// ***********************************************************************
// 
// Moocho: Multi-functional Object-Oriented arCHitecture for Optimization
//                  Copyright (2003) Sandia Corporation
// 
// Under terms of Contract DE-AC04-94AL85000, there is a non-exclusive
// license for use of this work by or on behalf of the U.S. Government.
// 
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//  
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
// Lesser General Public License for more details.
//  
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
// USA
// Questions? Contact Roscoe A. Bartlett (rabartl@sandia.gov) 
// 
// ***********************************************************************
// @HEADER

#include <ostream>

#include "MoochoPack_CalcLambdaIndepStd_AddedStep.hpp"
#include "MoochoPack_moocho_algo_conversion.hpp"
#include "IterationPack_print_algorithm_step.hpp"
#include "ConstrainedOptPack_ComputeMinMult.hpp"
#include "ConstrainedOptPack/src/VectorWithNorms.h"
#include "AbstractLinAlgPack_SpVectorOp.hpp"
#include "AbstractLinAlgPack/src/AbstractLinAlgPack_MatrixOp.hpp"
#include "DenseLinAlgPack_LinAlgOpPack.hpp"
#include "DenseLinAlgPack_DVectorOp.hpp"
#include "DenseLinAlgPack_DVectorOut.hpp"

namespace LinAlgOpPack {
  using AbstractLinAlgPack::Vp_StV;
  using AbstractLinAlgPack::Vp_StMtV;
}

bool MoochoPack::CalcLambdaIndepStd_AddedStep::do_step(Algorithm& _algo
  , poss_type step_poss, IterationPack::EDoStepType type, poss_type assoc_step_poss)
{

  using BLAS_Cpp::no_trans;
  using BLAS_Cpp::trans;

  using DenseLinAlgPack::Vt_S;
  using DenseLinAlgPack::norm_inf;

  using AbstractLinAlgPack::Vp_StMtV;
  
  using ConstrainedOptPack::min_abs;

  using LinAlgOpPack::Vp_V;
  using LinAlgOpPack::V_StMtV;
  using LinAlgOpPack::Vp_MtV;

  NLPAlgo &algo = rsqp_algo(_algo);
  NLPAlgoState  &s    = algo.rsqp_state();
  Range1D   decomp  = s.equ_decomp();
  
  EJournalOutputLevel olevel = algo.algo_cntr().journal_output_level();
  std::ostream& out = algo.track().journal_out();

  // print step header.
  if( static_cast<int>(olevel) >= static_cast<int>(PRINT_ALGORITHM_STEPS) ) {
    using IterationPack::print_algorithm_step;
    print_algorithm_step( algo, step_poss, type, assoc_step_poss, out );
  }

  // Compute: lambda(decomp) = inv(Gc(decomp)'* Y)' * ( - Y' * (Gf + nu) - U' * lambda(undecomp) )
  // where U = Gc(undecomp)' * Y

  // Must resize lambda here explicitly since we will only be updating a region of it.
  // If lambda(undecomp) has already been updated then lambda will have been resized
  // already but lambda(decomp) will not be initialized yet.
  if( !s.lambda().updated_k(0) ) s.lambda().set_k(0).v().resize( algo.nlp().m() );
  
  DVectorSlice lambda_decomp = s.lambda().get_k(0).v()(decomp);
  
  // lambda_decomp_tmp1 = - Y' * (Gf + nu)
  if( algo.nlp().has_bounds() ) {
    // _tmp = Gf + nu
    DVector _tmp = s.Gf().get_k(0)();
    DVectorSlice _vs_tmp = _tmp;  // only create this DVectorSlice once
    Vp_V( &_vs_tmp, s.nu().get_k(0)() );
    // lambda_decomp_tmp1 = - Y' * _tmp
    V_StMtV( &lambda_decomp, -1.0, s.Y().get_k(0), trans, _vs_tmp );
  }
  else {
    // lambda_decomp__tmp1 = - Y' * Gf
    V_StMtV( &lambda_decomp, -1.0, s.Y().get_k(0), trans, s.Gf().get_k(0)() );
  }

  // lambda_decomp_tmp2 = lambda_decomp_tmp1 - U' * lambda(undecomp)
  if( algo.nlp().r() < algo.nlp().m() ) {
    Range1D undecomp = s.equ_undecomp();
    Vp_StMtV( &lambda_decomp, -1.0, s.U().get_k(0), trans, s.lambda().get_k(0).v()(undecomp) );
  }
  // else lambda(decomp)_tmp2 = lambda(decomp)_tmp1

  // lambda_decomp = inv(Gc(decomp)'* Y)' * lambda_decomp_tmp2
  s.decomp_sys().solve_transAtY( lambda_decomp, trans, &lambda_decomp );

  if( static_cast<int>(olevel) >= static_cast<int>(PRINT_ALGORITHM_STEPS) ) {
    out << "\nmax(|lambda_k(equ_decomp)(i)|) = " << norm_inf(lambda_decomp)
      << "\nmin(|lambda_k(equ_decomp)(i)|) = " << min_abs(lambda_decomp)  << std::endl;
  }

  if( static_cast<int>(olevel) >= static_cast<int>(PRINT_VECTORS) ) {
    out << "\nlambda_k(equ_decomp) = \n" << lambda_decomp;
  }

  return true;
}

void MoochoPack::CalcLambdaIndepStd_AddedStep::print_step( const Algorithm& algo
  , poss_type step_poss, IterationPack::EDoStepType type, poss_type assoc_step_poss
  , std::ostream& out, const std::string& L ) const
{
  out
    << L << "*** Calculate the Lagrange multipliers for the decomposed constraints\n"
    << L << "lambda_k(equ_decomp) = - inv(Gc_k(:,equ_decomp)'*Y_k)\n"
    << L << "                        * (Y_k'*(Gf_k + nu_k) + U_k'*lambda_k(equ_undecomp))\n";
}

#endif // 0