MoochoPack_MoochoTrackerStatsStd.cpp

// @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.
// 
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are
// met:
//
// 1. Redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer.
//
// 2. Redistributions in binary form must reproduce the above copyright
// notice, this list of conditions and the following disclaimer in the
// documentation and/or other materials provided with the distribution.
//
// 3. Neither the name of the Corporation nor the names of the
// contributors may be used to endorse or promote products derived from
// this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY SANDIA CORPORATION "AS IS" AND ANY
// EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
// PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL SANDIA CORPORATION OR THE
// CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
// EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
// PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
// LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
// NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
// SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
//
// Questions? Contact Roscoe A. Bartlett (rabartl@sandia.gov) 
// 
// ***********************************************************************
// @HEADER

#include <assert.h>

#include <iomanip>

#include "MoochoPack_MoochoTrackerStatsStd.hpp"
#include "MoochoPack_NLPAlgoState.hpp"
#include "MoochoPack_moocho_algo_conversion.hpp"
#include "NLPInterfacePack_NLPFirstOrder.hpp"
#include "AbstractLinAlgPack_Vector.hpp"
#include "Teuchos_dyn_cast.hpp"

namespace {
template< class T >
inline
T my_max( const T& v1, const T& v2 ) { return v1 > v2 ? v1 : v2; }
} // end namespace

namespace MoochoPack {

using std::endl;
using std::setw;
using std::left;
using std::right;
using std::setprecision;

MoochoTrackerStatsStd::MoochoTrackerStatsStd(
  const ostream_ptr_t& o, const ostream_ptr_t& journal_out
  )
  :AlgorithmTracker(journal_out)
{
  set_output_stream(o);
}

void MoochoTrackerStatsStd::set_output_stream(const ostream_ptr_t& o)
{
  o_ = o;
}

const MoochoTrackerStatsStd::ostream_ptr_t&
MoochoTrackerStatsStd::get_output_stream() const
{
  return o_;
}

void MoochoTrackerStatsStd::initialize()
{
  num_QN_updates_ = 0;
  timer_.reset();
  timer_.start();
}

void MoochoTrackerStatsStd::output_iteration(const Algorithm& p_algo) const
{
  const NLPAlgo  &algo = rsqp_algo(p_algo);
  const NLPAlgoState &s    = algo.rsqp_state();

  // All we have to do here is to just to count the number of quasi-newton updates
  const QuasiNewtonStats  *quasi_newt_stats =
    ( quasi_newton_stats_.exists_in(s) && quasi_newton_stats_(s).updated_k(0)
      ? &quasi_newton_stats_(s).get_k(0)
      : NULL );
  if( quasi_newt_stats ) {
    QuasiNewtonStats::EUpdate updated = quasi_newt_stats->updated();
    if( updated == QuasiNewtonStats::DAMPENED_UPDATED || updated == QuasiNewtonStats::UPDATED )
      num_QN_updates_++;
  }
}

void MoochoTrackerStatsStd::output_final( const Algorithm& p_algo
  , EAlgoReturn algo_return ) const
{
  using Teuchos::dyn_cast;

  const NLPAlgo           &algo    = rsqp_algo(p_algo);
  const NLPAlgoState          &s       = algo.rsqp_state();
  const NLPObjGrad     &nlp     = dyn_cast<const NLPObjGrad>(algo.nlp()); 
  const NLPFirstOrder  *nlp_foi = dynamic_cast<const NLPFirstOrder*>(&nlp); 

  const size_type
    m = nlp.m();

  std::ostream& o = this->o();

  // Stop the timer
  timer_.stop();

  // Formating info
  const int
    p      = 18,
    stat_w = 15,
    val_w  = p + 10;

  // Get a Quasi-Newton statistics.
  const QuasiNewtonStats  *quasi_newt_stats =
    ( quasi_newton_stats_.exists_in(s) && quasi_newton_stats_(s).updated_k(0)
      ? &quasi_newton_stats_(s).get_k(0)
      : NULL );
  if( quasi_newt_stats ) {
    QuasiNewtonStats::EUpdate updated = quasi_newt_stats->updated();
    if( updated == QuasiNewtonStats::DAMPENED_UPDATED || updated == QuasiNewtonStats::UPDATED )
      num_QN_updates_++;
  }

  // status
  o << left << setw(stat_w) << "status" << "= "
    << right << setw(val_w);
  switch( algo_return ) {
    case IterationPack::TERMINATE_TRUE:
      o << "solved";
      break;
    case IterationPack::TERMINATE_FALSE:
      o << "except";
      break;
    case IterationPack::MAX_ITER_EXCEEDED:
      o << "max_iter";
      break;
    case IterationPack::MAX_RUN_TIME_EXCEEDED:
      o << "max_run_time";
      break;
    case IterationPack::INTERRUPTED_TERMINATE_TRUE:
      o << "interrupted_solved";
      break;
    case IterationPack::INTERRUPTED_TERMINATE_FALSE:
      o << "interrupted_not_solved";
      break;
    default:
      TEUCHOS_TEST_FOR_EXCEPT(true);
  }
  o << "; # solved, except, max_iter, max_run_time\n";
  // niter
  o << left << setw(stat_w) << "niter" << "= "
    << right << setw(val_w) << s.k()
    << "; # Number of rSQP iterations (plus 1?)\n";
  // nfunc
  o << left << setw(stat_w) << "nfunc" << "= "
    << right << setw(val_w) << my_max(nlp.num_f_evals(),(m? nlp.num_c_evals():0) )
    << "; # max( number f(x) evals, number c(x) evals )\n";
  // ngrad
  o << left << setw(stat_w) << "ngrad" << "= "
    << right << setw(val_w) << my_max(nlp.num_Gf_evals(),(m?(nlp_foi?nlp_foi->num_Gc_evals():s.k()+1):0))
    << "; # max( number Gf(x) evals, number Gc(x) evals )\n";
  // CPU
  o << left << setw(stat_w) << "CPU" << "= "
    << right << setw(val_w) << timer_.read()
    << "; # Number of CPU seconds total\n";
  // obj_func
  o << left << setw(stat_w) << "obj_func" << "= "
    << right << setw(val_w);
  if(s.f().updated_k(0))
    o << s.f().get_k(0);
  else
    o << "-";
  o << "; # Objective function value f(x) at final point\n";
  // feas_kkt_err
  o << left << setw(stat_w) << "feas_kkt_err" << "= "
    << right << setw(val_w);
  if(s.feas_kkt_err().updated_k(0))
    o << s.feas_kkt_err().get_k(0);
  else if(s.c().updated_k(0))
    o << s.c().get_k(0).norm_inf();
  else
    o << "-";
  o << "; # Feasibility error at final point (scaled ||c(x)||inf, feas_err_k)\n";
  // opt_kkt_err
  o << left << setw(stat_w) << "opt_kkt_err" << "= "
    << right << setw(val_w);
  if(s.opt_kkt_err().updated_k(0))
    o << s.opt_kkt_err().get_k(0);
  else if(s.rGL().updated_k(0))
    o << s.rGL().get_k(0).norm_inf();
  else if(s.rGL().updated_k(-1))
    o << s.rGL().get_k(-1).norm_inf();
  else
    o << "-";
  o << "; # Optimality error at final point (scaled ||rGL||inf, opt_err_k)\n";
  // nact
  o << left << setw(stat_w) << "nact" << "= "
    << right << setw(val_w);
  if(s.nu().updated_k(0))
    o << s.nu().get_k(0).nz();
  else if(s.nu().updated_k(-1))
    o << s.nu().get_k(-1).nz();
  else
    o << "-";
  o << "; # Number of total active constraints at the final point\n";
  // nbasis_change
  const IterQuantityAccess<index_type> &num_basis = s.num_basis();
  const int lu_k = num_basis.last_updated();
  o << left << setw(stat_w) << "nbasis_change" << "= "
    << right << setw(val_w) << ( lu_k != IterQuantity::NONE_UPDATED
                   ? num_basis.get_k(lu_k)
                   : 0 ) 
    << "; # Number of basis changes\n";
  // nquasi_newton
  o << left << setw(stat_w) << "nquasi_newton" << "= "
    << right << setw(val_w) << num_QN_updates_
    << "; # Number of quasi-newton updates\n";

}

} // end namespace MoochoPack