ROL_OptimizationProblem.hpp

Go to the documentation of this file.

// @HEADER
// ************************************************************************
//
//               Rapid Optimization Library (ROL) Package
//                 Copyright (2014) 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 lead developers:
//              Drew Kouri   (dpkouri@sandia.gov) and
//              Denis Ridzal (dridzal@sandia.gov)
//
// ************************************************************************
// @HEADER

#ifndef ROL_OPTIMIZATIONPROBLEM_HPP
#define ROL_OPTIMIZATIONPROBLEM_HPP

#include "ROL_ConstraintManager.hpp"
#include "ROL_SlacklessObjective.hpp"

// Stochastic Includes
#include "ROL_SampleGenerator.hpp"
#include "ROL_RiskVector.hpp"
#include "ROL_RiskBoundConstraint.hpp"
// Objective includes
#include "ROL_RiskNeutralObjective.hpp" 
#include "ROL_StochasticObjective.hpp"
#include "ROL_RiskLessObjective.hpp"
// Constraint includes
#include "ROL_RiskNeutralConstraint.hpp" 
#include "ROL_StochasticConstraint.hpp" 
#include "ROL_RiskLessConstraint.hpp"
// Almost sure constraint includes
#include "ROL_AlmostSureConstraint.hpp"
#include "ROL_SimulatedBoundConstraint.hpp"
#include "ROL_SimulatedVector.hpp"

namespace ROL {

/* Represents optimization problems in Type-EB form 
 */

template<class Real>
class OptimizationProblem {
private:
  Ptr<Objective<Real>>                    INPUT_obj_;
  Ptr<Vector<Real>>                       INPUT_sol_;
  Ptr<BoundConstraint<Real>>              INPUT_bnd_;
  std::vector<Ptr<Constraint<Real>>>      INPUT_econ_;
  std::vector<Ptr<Vector<Real>>>          INPUT_emul_;
  std::vector<Ptr<Constraint<Real>>>      INPUT_icon_;
  std::vector<Ptr<Vector<Real>>>          INPUT_imul_;
  std::vector<Ptr<BoundConstraint<Real>>> INPUT_ibnd_;

  Ptr<Objective<Real>>                    INTERMEDIATE_obj_;
  Ptr<Vector<Real>>                       INTERMEDIATE_sol_;
  Ptr<BoundConstraint<Real>>              INTERMEDIATE_bnd_;
  std::vector<Ptr<Constraint<Real>>>      INTERMEDIATE_econ_;
  std::vector<Ptr<Vector<Real>>>          INTERMEDIATE_emul_;
  std::vector<Ptr<Constraint<Real>>>      INTERMEDIATE_icon_;
  std::vector<Ptr<Vector<Real>>>          INTERMEDIATE_imul_;
  std::vector<Ptr<BoundConstraint<Real>>> INTERMEDIATE_ibnd_;

  Ptr<SampleGenerator<Real>>              vsampler_;
  Ptr<SampleGenerator<Real>>              gsampler_;
  Ptr<SampleGenerator<Real>>              hsampler_;
  std::vector<Ptr<SampleGenerator<Real>>> exsampler_;
  std::vector<Ptr<BatchManager<Real>>>    ecbman_;
  std::vector<Ptr<SampleGenerator<Real>>> ixsampler_;
  std::vector<Ptr<BatchManager<Real>>>    icbman_;

  Ptr<ParameterList>                      parlistObj_;
  std::vector<Ptr<ParameterList>>         parlistCon_;

  Ptr<Objective<Real>>       obj_;
  Ptr<Vector<Real>>          sol_;
  Ptr<BoundConstraint<Real>> bnd_;
  Ptr<Constraint<Real>>      con_;
  Ptr<Vector<Real>>          mul_;

  Ptr<ConstraintManager<Real>> conManager_;

  EProblem problemType_;

  bool isInitialized_;

  bool              needRiskLessObj_;
  std::vector<bool> needRiskLessEcon_;
  std::vector<bool> needRiskLessIcon_;
  bool              isStochastic_;

  void initialize( const Ptr<Objective<Real>>                    &obj,
                   const Ptr<Vector<Real>>                       &x,
                   const Ptr<BoundConstraint<Real>>              &bnd,
                   const std::vector<Ptr<Constraint<Real>>>      &econ,
                   const std::vector<Ptr<Vector<Real>>>          &emul,
                   const std::vector<Ptr<Constraint<Real>>>      &icon,
                   const std::vector<Ptr<Vector<Real>>>          &imul,
                   const std::vector<Ptr<BoundConstraint<Real>>> &ibnd ) {
    if (!isInitialized_) {
      int esize = static_cast<int>(econ.size());
      int isize = static_cast<int>(icon.size());
      std::vector<Ptr<Constraint<Real>>>      cvec;
      std::vector<Ptr<Vector<Real>>>          lvec;
      std::vector<Ptr<BoundConstraint<Real>>> bvec;
      for (int i = 0; i < esize; ++i) {
        if ( econ[i] != nullPtr ) {
          if (isStochastic_) {
            cvec.push_back(setRiskLessCon(econ[i],needRiskLessEcon_[i]));
          }
          else {
            cvec.push_back(econ[i]);
          }
          lvec.push_back(emul[i]);
          bvec.push_back(nullPtr);
        }
      }
      for (int i = 0; i < isize; ++i) {
        if ( icon[i] != nullPtr ) {
          if (isStochastic_) {
            cvec.push_back(setRiskLessCon(icon[i],needRiskLessIcon_[i]));
          }
          else {
            cvec.push_back(icon[i]);
          }
          lvec.push_back(imul[i]);
          bvec.push_back(ibnd[i]);
        }
      }
 
      conManager_ = makePtr<ConstraintManager<Real>>(cvec,lvec,bvec,x,bnd);
      con_        = conManager_->getConstraint();
      mul_        = conManager_->getMultiplier();
      sol_        = conManager_->getOptVector();
      bnd_        = conManager_->getBoundConstraint();
      Ptr<Objective<Real>> obj0;
      if (isStochastic_) {
        obj0 = setRiskLessObj(obj,needRiskLessObj_);
      }
      else {
        obj0 = obj;
      }
      if ( conManager_->hasInequality() ) {
        obj_      = makePtr<SlacklessObjective<Real>>(obj0);
      }
      else {
        obj_      = obj0;
      }

      if ( conManager_->isNull() ) {
        if( bnd_ == nullPtr || !bnd_->isActivated() ) {  // Type-U
          problemType_ = TYPE_U;        
        }
        else { // Type-B
          problemType_ = TYPE_B; 
        }
      }
      else {
        if( bnd_ == nullPtr || !bnd_->isActivated() ) { // Type-E
          problemType_ = TYPE_E;     
        }
        else { // Type-EB
          problemType_ = TYPE_EB; 
        }
      }
      isInitialized_ = true;
    }
  }

  const Ptr<Constraint<Real>> setRiskLessCon(const Ptr<Constraint<Real>> &con, const bool needRiskLess) const {
    if (needRiskLess) {
      return makePtr<RiskLessConstraint<Real>>(con);
    }
    else {
      return con;
    }
  }

  const Ptr<Objective<Real>> setRiskLessObj(const Ptr<Objective<Real>> &obj, const bool needRiskLess) const {
    if (needRiskLess) {
      return makePtr<RiskLessObjective<Real>>(obj);
    }
    else {
      return obj;
    }
  }

  std::vector<Real> computeSampleMean(const Ptr<SampleGenerator<Real>> &sampler) const {
    // Compute mean value of inputs and set parameter in objective
    int dim = sampler->getMyPoint(0).size(), nsamp = sampler->numMySamples();
    std::vector<Real> loc(dim), mean(dim), pt(dim);
    Real wt(0);
    for (int i = 0; i < nsamp; i++) {
      pt = sampler->getMyPoint(i);
      wt = sampler->getMyWeight(i);
      for (int j = 0; j < dim; j++) {
        loc[j] += wt*pt[j];
      }
    }
    sampler->sumAll(&loc[0],&mean[0],dim);
    return mean;
  }

  void initStochastic(void) {
    if (!isStochastic_) {
      int econSize = INPUT_econ_.size();
      int iconSize = INPUT_icon_.size();
      needRiskLessObj_ = true;
      needRiskLessEcon_.clear(); needRiskLessEcon_.resize(econSize,true);
      needRiskLessIcon_.clear(); needRiskLessIcon_.resize(iconSize,true);
      parlistObj_ = nullPtr;
      parlistCon_.clear(); parlistCon_.resize(iconSize,nullPtr);

      exsampler_.clear(); exsampler_.resize(econSize,nullPtr);
      ecbman_.clear();    ecbman_.resize(econSize,nullPtr);

      ixsampler_.clear(); ixsampler_.resize(iconSize,nullPtr);
      icbman_.clear();    icbman_.resize(iconSize,nullPtr);

      isStochastic_ = true;
    }
  }

  void buildRiskVec(Ptr<Vector<Real>> &x) {
    // Build risk vector and risk bound constraint
    INTERMEDIATE_sol_
      = makePtr<RiskVector<Real>>(parlistObj_,parlistCon_,x);
    if (parlistObj_ != nullPtr) {
      Real statObj = parlistObj_->sublist("SOL").get("Initial Statistic",1.0);
      dynamicPtrCast<RiskVector<Real>>(INTERMEDIATE_sol_)->setStatistic(statObj,0);
    }
    int nc = INPUT_icon_.size();
    for (int i = 0; i < nc; ++i) {
      if (parlistCon_[i] != nullPtr) {
        Real statCon = parlistCon_[i]->sublist("SOL").get("Initial Statistic",1.0);
        dynamicPtrCast<RiskVector<Real>>(INTERMEDIATE_sol_)->setStatistic(statCon,1,i);
      }
    }
  }

  void buildRiskBnd(Ptr<BoundConstraint<Real>> &bnd) {
    if ( INPUT_bnd_ != nullPtr ) {
      INTERMEDIATE_bnd_
        = makePtr<RiskBoundConstraint<Real>>(parlistObj_,parlistCon_,bnd);
    }
  }

public:
  virtual ~OptimizationProblem(void) {}

  // Default constructor [0]
  OptimizationProblem(void)
   : isInitialized_(false), isStochastic_(false) {}

  // Complete option constructor [1]
  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const Ptr<BoundConstraint<Real>>              &bnd,
                       const std::vector<Ptr<Constraint<Real>>>      &econ,
                       const std::vector<Ptr<Vector<Real>>>          &emul,
                       const std::vector<Ptr<Constraint<Real>>>      &icon,
                       const std::vector<Ptr<Vector<Real>>>          &imul,
                       const std::vector<Ptr<BoundConstraint<Real>>> &ibnd )
    : INPUT_obj_(obj), INPUT_sol_(x), INPUT_bnd_(bnd),
      INPUT_econ_(econ), INPUT_emul_(emul),
      INPUT_icon_(icon), INPUT_imul_(imul), INPUT_ibnd_(ibnd),
      INTERMEDIATE_obj_(obj), INTERMEDIATE_sol_(x), INTERMEDIATE_bnd_(bnd),
      INTERMEDIATE_econ_(econ), INTERMEDIATE_emul_(emul),
      INTERMEDIATE_icon_(icon), INTERMEDIATE_imul_(imul), INTERMEDIATE_ibnd_(ibnd),
      isInitialized_(false), isStochastic_(false) {}

  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const Ptr<BoundConstraint<Real>>              &bnd,
                       const Ptr<Constraint<Real>>                   &econ,
                       const Ptr<Vector<Real>>                       &emul,
                       const std::vector<Ptr<Constraint<Real>>>      &icon,
                       const std::vector<Ptr<Vector<Real>>>          &imul,
                       const std::vector<Ptr<BoundConstraint<Real>>> &ibnd )
    : INPUT_obj_(obj), INPUT_sol_(x), INPUT_bnd_(bnd),
      INPUT_icon_(icon), INPUT_imul_(imul), INPUT_ibnd_(ibnd),
      INTERMEDIATE_obj_(obj), INTERMEDIATE_sol_(x), INTERMEDIATE_bnd_(bnd),
      INTERMEDIATE_icon_(icon), INTERMEDIATE_imul_(imul), INTERMEDIATE_ibnd_(ibnd),
      isInitialized_(false), isStochastic_(false) {
    std::vector<Ptr<Constraint<Real>>> econ0(1,econ);
    std::vector<Ptr<Vector<Real>>>     emul0(1,emul);
    INPUT_econ_ = econ0;
    INPUT_emul_ = emul0;
    INTERMEDIATE_econ_ = econ0;
    INTERMEDIATE_emul_ = emul0;
  }

  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const Ptr<BoundConstraint<Real>>              &bnd,
                       const std::vector<Ptr<Constraint<Real>>>      &econ,
                       const std::vector<Ptr<Vector<Real>>>          &emul,
                       const Ptr<Constraint<Real>>                   &icon,
                       const Ptr<Vector<Real>>                       &imul,
                       const Ptr<BoundConstraint<Real>>              &ibnd )
    : INPUT_obj_(obj), INPUT_sol_(x), INPUT_bnd_(bnd),
      INPUT_econ_(econ), INPUT_emul_(emul),
      INTERMEDIATE_obj_(obj), INTERMEDIATE_sol_(x), INTERMEDIATE_bnd_(bnd),
      INTERMEDIATE_econ_(econ), INTERMEDIATE_emul_(emul),
      isInitialized_(false), isStochastic_(false) {
    std::vector<Ptr<Constraint<Real>>>      icon0(1,icon);
    std::vector<Ptr<Vector<Real>>>          imul0(1,imul);
    std::vector<Ptr<BoundConstraint<Real>>> ibnd0(1,ibnd);
    INPUT_icon_ = icon0;
    INPUT_imul_ = imul0;
    INPUT_ibnd_ = ibnd0;
    INTERMEDIATE_icon_ = icon0;
    INTERMEDIATE_imul_ = imul0;
    INTERMEDIATE_ibnd_ = ibnd0;
  }

  OptimizationProblem( const Ptr<Objective<Real>>                     &obj,
                       const Ptr<Vector<Real>>                        &x,
                       const Ptr<BoundConstraint<Real>>               &bnd,
                       const Ptr<Constraint<Real>>                    &econ,
                       const Ptr<Vector<Real>>                        &emul,
                       const Ptr<Constraint<Real>>                    &icon,
                       const Ptr<Vector<Real>>                        &imul,
                       const Ptr<BoundConstraint<Real>>               &ibnd )
    : INPUT_obj_(obj), INPUT_sol_(x), INPUT_bnd_(bnd),
      INTERMEDIATE_obj_(obj), INTERMEDIATE_sol_(x), INTERMEDIATE_bnd_(bnd),
      isInitialized_(false), isStochastic_(false) {
    std::vector<Ptr<Constraint<Real>>>      econ0(1,econ);
    std::vector<Ptr<Vector<Real>>>          emul0(1,emul);
    std::vector<Ptr<Constraint<Real>>>      icon0(1,icon);
    std::vector<Ptr<Vector<Real>>>          imul0(1,imul);
    std::vector<Ptr<BoundConstraint<Real>>> ibnd0(1,ibnd);
    INPUT_econ_ = econ0;
    INPUT_emul_ = emul0;
    INPUT_icon_ = icon0;
    INPUT_imul_ = imul0;
    INPUT_ibnd_ = ibnd0;
    INTERMEDIATE_econ_ = econ0;
    INTERMEDIATE_emul_ = emul0;
    INTERMEDIATE_icon_ = icon0;
    INTERMEDIATE_imul_ = imul0;
    INTERMEDIATE_ibnd_ = ibnd0;
  }

  // No bound constuctor [2]
  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const std::vector<Ptr<Constraint<Real>>>      &econ,
                       const std::vector<Ptr<Vector<Real>>>          &emul,
                       const std::vector<Ptr<Constraint<Real>>>      &icon,
                       const std::vector<Ptr<Vector<Real>>>          &imul,
                       const std::vector<Ptr<BoundConstraint<Real>>> &ibnd )
    : OptimizationProblem( obj, x, nullPtr, econ, emul, icon, imul, ibnd ) {}

  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const Ptr<Constraint<Real>>                   &econ,
                       const Ptr<Vector<Real>>                       &emul,
                       const std::vector<Ptr<Constraint<Real>>>      &icon,
                       const std::vector<Ptr<Vector<Real>>>          &imul,
                       const std::vector<Ptr<BoundConstraint<Real>>> &ibnd )
    : OptimizationProblem( obj, x, nullPtr, econ, emul, icon, imul, ibnd) {}

  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const std::vector<Ptr<Constraint<Real>>>      &econ,
                       const std::vector<Ptr<Vector<Real>>>          &emul,
                       const Ptr<Constraint<Real>>                   &icon,
                       const Ptr<Vector<Real>>                       &imul,
                       const Ptr<BoundConstraint<Real>>              &ibnd )
    : OptimizationProblem( obj, x, nullPtr, econ, emul, icon, imul, ibnd) {}

  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const Ptr<Constraint<Real>>                   &econ,
                       const Ptr<Vector<Real>>                       &emul,
                       const Ptr<Constraint<Real>>                   &icon,
                       const Ptr<Vector<Real>>                       &imul,
                       const Ptr<BoundConstraint<Real>>              &ibnd )
    : OptimizationProblem( obj, x, nullPtr, econ, emul, icon, imul, ibnd) {}

  // No inequality constraint [3]
  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const Ptr<BoundConstraint<Real>>              &bnd,
                       const std::vector<Ptr<Constraint<Real>>>      &econ,
                       const std::vector<Ptr<Vector<Real>>>          &emul )
    : OptimizationProblem( obj, x, bnd, econ, emul, nullPtr, nullPtr, nullPtr ) {}

  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const Ptr<BoundConstraint<Real>>              &bnd,
                       const Ptr<Constraint<Real>>                   &econ,
                       const Ptr<Vector<Real>>                       &emul )
    : OptimizationProblem( obj, x, bnd, econ, emul, nullPtr, nullPtr, nullPtr) {}

  // No equality constraint [4]
  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const Ptr<BoundConstraint<Real>>              &bnd,
                       const std::vector<Ptr<Constraint<Real>>>      &icon,
                       const std::vector<Ptr<Vector<Real>>>          &imul,
                       const std::vector<Ptr<BoundConstraint<Real>>> &ibnd )
    : OptimizationProblem( obj, x, bnd, nullPtr, nullPtr, icon, imul, ibnd ) {}

  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const Ptr<BoundConstraint<Real>>              &bnd,
                       const Ptr<Constraint<Real>>                   &icon,
                       const Ptr<Vector<Real>>                       &imul,
                       const Ptr<BoundConstraint<Real>>              &ibnd )
    : OptimizationProblem( obj, x, bnd, nullPtr, nullPtr, icon, imul, ibnd) {}

  // No inequality or bound constraint [5]
  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const std::vector<Ptr<Constraint<Real>>>      &econ,
                       const std::vector<Ptr<Vector<Real>>>          &emul )
    : OptimizationProblem( obj, x, nullPtr, econ, emul, nullPtr, nullPtr, nullPtr ) {}

  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const Ptr<Constraint<Real>>                   &econ,
                       const Ptr<Vector<Real>>                       &emul )
    : OptimizationProblem( obj, x, nullPtr, econ, emul, nullPtr, nullPtr, nullPtr) {}

  // No equality or bound constraint [6]
  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const std::vector<Ptr<Constraint<Real>>>      &icon,
                       const std::vector<Ptr<Vector<Real>>>          &imul,
                       const std::vector<Ptr<BoundConstraint<Real>>> &ibnd )
    : OptimizationProblem( obj, x, nullPtr, nullPtr, nullPtr, icon, imul, ibnd ) {}

  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const Ptr<Constraint<Real>>                   &icon,
                       const Ptr<Vector<Real>>                       &imul,
                       const Ptr<BoundConstraint<Real>>              &ibnd )
    : OptimizationProblem( obj, x, nullPtr, nullPtr, nullPtr, icon, imul, ibnd) {}

  // Bound constrained problem [7]
  OptimizationProblem( const Ptr<Objective<Real>>                    &obj,
                       const Ptr<Vector<Real>>                       &x,
                       const Ptr<BoundConstraint<Real>>              &bnd )
    : OptimizationProblem( obj, x, bnd, nullPtr, nullPtr, nullPtr, nullPtr, nullPtr ) {}

  // Unconstrained problem [8]
  OptimizationProblem( const Ptr<Objective<Real>> &obj,
                       const Ptr<Vector<Real>>    &x ) :
     OptimizationProblem( obj, x, nullPtr, nullPtr, nullPtr, nullPtr, nullPtr, nullPtr ) {} 

  /* Get methods */

  virtual Ptr<Objective<Real>> getObjective(void) {
    if ( INTERMEDIATE_obj_ == nullPtr ) {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::getObjective: No objective inputed!");
    }
    initialize(INTERMEDIATE_obj_,INTERMEDIATE_sol_,INTERMEDIATE_bnd_,
               INTERMEDIATE_econ_,INTERMEDIATE_emul_,
               INTERMEDIATE_icon_,INTERMEDIATE_imul_,INTERMEDIATE_ibnd_);
    return obj_;
  }

  virtual Ptr<Vector<Real>> getSolutionVector(void) {
    if ( INTERMEDIATE_sol_ == nullPtr ) {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::getSolutionVector: No solution vector inputed!");
    }
    initialize(INTERMEDIATE_obj_,INTERMEDIATE_sol_,INTERMEDIATE_bnd_,
               INTERMEDIATE_econ_,INTERMEDIATE_emul_,
               INTERMEDIATE_icon_,INTERMEDIATE_imul_,INTERMEDIATE_ibnd_);
    return sol_;
  }

  virtual Ptr<BoundConstraint<Real>> getBoundConstraint(void) {
    initialize(INTERMEDIATE_obj_,INTERMEDIATE_sol_,INTERMEDIATE_bnd_,
               INTERMEDIATE_econ_,INTERMEDIATE_emul_,
               INTERMEDIATE_icon_,INTERMEDIATE_imul_,INTERMEDIATE_ibnd_);
    return bnd_;
  }

  virtual Ptr<Constraint<Real>> getConstraint(void) {
    initialize(INTERMEDIATE_obj_,INTERMEDIATE_sol_,INTERMEDIATE_bnd_,
               INTERMEDIATE_econ_,INTERMEDIATE_emul_,
               INTERMEDIATE_icon_,INTERMEDIATE_imul_,INTERMEDIATE_ibnd_);
    return con_;
  }

  virtual Ptr<Vector<Real>> getMultiplierVector(void) {
    initialize(INTERMEDIATE_obj_,INTERMEDIATE_sol_,INTERMEDIATE_bnd_,
               INTERMEDIATE_econ_,INTERMEDIATE_emul_,
               INTERMEDIATE_icon_,INTERMEDIATE_imul_,INTERMEDIATE_ibnd_);
    return mul_;
  }

  EProblem getProblemType(void) {
    initialize(INTERMEDIATE_obj_,INTERMEDIATE_sol_,INTERMEDIATE_bnd_,
               INTERMEDIATE_econ_,INTERMEDIATE_emul_,
               INTERMEDIATE_icon_,INTERMEDIATE_imul_,INTERMEDIATE_ibnd_);
    return problemType_;
  }

  /* Set Stochastic Methods */

  /* Objective function */
  void setMeanValueObjective(const Ptr<SampleGenerator<Real>> &sampler) {
    initStochastic();
    // Set objective function samplers
    vsampler_ = sampler;
    gsampler_ = sampler;
    hsampler_ = sampler;
    // Construct risk-averse/probabilistic objective function
    if ( vsampler_ == nullPtr ) {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::setMeanValueObjective: Objective function value sampler is null!");
    }
    else {
      std::vector<Real> mean = computeSampleMean(vsampler_);
      INTERMEDIATE_obj_ = INPUT_obj_;
      INTERMEDIATE_obj_->setParameter(mean);
    }
    // Set vector and bound constraint
    buildRiskVec(INPUT_sol_);
    buildRiskBnd(INPUT_bnd_);

    isInitialized_ = false;
  }

  void setRiskNeutralObjective(const Ptr<SampleGenerator<Real>> &vsampler,
                               const Ptr<SampleGenerator<Real>> &gsampler = nullPtr,
                               const Ptr<SampleGenerator<Real>> &hsampler = nullPtr,
                               const bool storage = true) {
    initStochastic();
    // Set objective function samplers
    vsampler_ = vsampler;
    gsampler_ = gsampler;
    hsampler_ = hsampler;
    if ( gsampler == nullPtr ) {
      gsampler_ = vsampler_;
    }
    if ( hsampler == nullPtr ) {
      hsampler_ = gsampler_;
    }
    // Construct risk-averse/probabilistic objective function
    if ( vsampler_ == nullPtr ) {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::setRiskNeutralObjective: Objective function value sampler is null!");
    }
    else {
      INTERMEDIATE_obj_
        = makePtr<RiskNeutralObjective<Real>>(INPUT_obj_,vsampler_,gsampler_,hsampler_,storage);
    }
    // Set vector and bound constraint
    buildRiskVec(INPUT_sol_);
    buildRiskBnd(INPUT_bnd_);

    isInitialized_ = false;
  }

  void setRiskAverseObjective(ParameterList &parlist,
                              const Ptr<SampleGenerator<Real>> &vsampler,
                              const Ptr<SampleGenerator<Real>> &gsampler = nullPtr,
                              const Ptr<SampleGenerator<Real>> &hsampler = nullPtr) {
    initStochastic();
    // Set objective function samplers
    vsampler_ = vsampler;
    gsampler_ = gsampler;
    hsampler_ = hsampler;
    if ( gsampler == nullPtr ) {
      gsampler_ = vsampler_;
    }
    if ( hsampler == nullPtr ) {
      hsampler_ = gsampler_;
    }
    // Construct risk-averse/probabilistic objective function
    if ( vsampler_ == nullPtr ) {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::setRiskAverseObjective: Objective function value sampler is null!");
    }
    else {
      needRiskLessObj_ = false;
      parlistObj_      = makePtrFromRef(parlist);
      INTERMEDIATE_obj_
        = makePtr<StochasticObjective<Real>>(INPUT_obj_,parlist,vsampler_,gsampler_,hsampler_);
    }
    // Set vector and bound constraint
    buildRiskVec(INPUT_sol_);
    buildRiskBnd(INPUT_bnd_);

    isInitialized_ = false;
  }

  void setStochasticObjective(ParameterList &parlist,
                              const Ptr<SampleGenerator<Real>> &vsampler,
                              const Ptr<SampleGenerator<Real>> &gsampler = nullPtr,
                              const Ptr<SampleGenerator<Real>> &hsampler = nullPtr) {
    // Determine Stochastic Objective Type
    std::string type = parlist.sublist("SOL").get("Stochastic Component Type","Risk Neutral");
    if ( type == "Risk Neutral" ) {
      bool storage = parlist.sublist("SOL").get("Store Sampled Value and Gradient",true);
      setRiskNeutralObjective(vsampler,gsampler,hsampler,storage);
    }
    else if ( type == "Risk Averse" ||
              type == "Deviation"   ||
              type == "Error"       ||
              type == "Regret"      ||
              type == "Probability" ) {
      setRiskAverseObjective(parlist,vsampler,gsampler,hsampler);
    }
    else if ( type == "Mean Value" ) {
      setMeanValueObjective(vsampler);
    }
    else {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::setStochasticObjective: Invalid stochastic optimization type!");
    }
    // Set vector and bound constraint
    buildRiskVec(INPUT_sol_);
    buildRiskBnd(INPUT_bnd_);

    isInitialized_ = false;
  }

  /* Equality Constraint */
  void setMeanValueEquality(const Ptr<SampleGenerator<Real>> &sampler, const int index = 0) {
    initStochastic();
    exsampler_[index] = sampler;
    if ( INPUT_econ_[index] != nullPtr && sampler != nullPtr ) {
      std::vector<Real> mean = computeSampleMean(sampler);
      INTERMEDIATE_econ_[index] = INPUT_econ_[index];
      INTERMEDIATE_econ_[index]->setParameter(mean);
      INTERMEDIATE_emul_[index] = INPUT_emul_[index];
    }
    else {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::setMeanValueEquality: Either SampleGenerator or Constraint is NULL!");
    }
    // Set vector and bound constraint
    buildRiskVec(INPUT_sol_);
    buildRiskBnd(INPUT_bnd_);

    isInitialized_ = false;
  }

  void setRiskNeutralEquality(const Ptr<SampleGenerator<Real>> &xsampler,
                              const Ptr<BatchManager<Real>>    &cbman,
                              const int index = 0) {
    initStochastic();
    exsampler_[index] = xsampler;
    ecbman_[index]    = cbman;
    if ( INPUT_econ_[index] != nullPtr
         &&        xsampler != nullPtr
         &&           cbman != nullPtr ) {
      INTERMEDIATE_econ_[index]
        = makePtr<RiskNeutralConstraint<Real>>(INPUT_econ_[index],xsampler,cbman);
      INTERMEDIATE_emul_[index] = INPUT_emul_[index];
    }
    else {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::SetRiskNeutralEquality: Either SampleGenerator, Constraint or BatchManager is NULL!");
    }
    // Set vector and bound constraint
    buildRiskVec(INPUT_sol_);
    buildRiskBnd(INPUT_bnd_);

    isInitialized_ = false;
  }

  void setAlmostSureEquality(const Ptr<SampleGenerator<Real>> &sampler,
                             const int index = 0) {
    initStochastic();
    exsampler_[index] = sampler;
    if ( INPUT_econ_[index] != nullPtr && sampler != nullPtr ) {
      int nsamp = sampler->numMySamples();
      INTERMEDIATE_econ_[index]
        = makePtr<AlmostSureConstraint<Real>>(sampler,INPUT_econ_[index]);
      std::vector<Ptr<Vector<Real>>> emul(nsamp,nullPtr);
      for (int j = 0; j < nsamp; ++j) {
        emul[j] = INPUT_emul_[index]->clone();
        emul[j]->set(*INPUT_emul_[index]);
      }
      INTERMEDIATE_emul_[index]
        = makePtr<DualSimulatedVector<Real>>(emul, sampler->getBatchManager(), sampler);
    }
    else {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::SetAlmostSureEquality: Either SampleGenerator or Constraint is NULL!");
    }
    // Set vector and bound constraint
    buildRiskVec(INPUT_sol_);
    buildRiskBnd(INPUT_bnd_);

    isInitialized_ = false;
  }


  void setStochasticEquality(std::vector<ParameterList> &parlist,
                             const std::vector<Ptr<SampleGenerator<Real>>> &xsampler,
                             const std::vector<Ptr<BatchManager<Real>>> &cbman) {
    initStochastic();
    int nc = static_cast<int>(INPUT_econ_.size());
    if ( nc != static_cast<int>(xsampler.size()) || nc != static_cast<int>(cbman.size()) ) {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::setStochasticEquality: Constraint vector and SampleGenerator vector are not the same size!");
    }
    for (int i = 0; i < nc; ++i) {
      if (xsampler[i] != nullPtr) {
        std::string type = parlist[i].sublist("SOL").get("Stochastic Component Type","Risk Neutral");
        if ( type == "Risk Neutral" ) {
          setRiskNeutralEquality(xsampler[i],cbman[i],i);
        }
        else if ( type == "Almost Sure" ) {
          setAlmostSureEquality(xsampler[i],i);
        }
        else if ( type == "Mean Value" ) {
          setMeanValueEquality(xsampler[i],i);
        }
        else {
          throw Exception::NotImplemented(">>> ROL::OptimizationProblem::SetStochasticEquality: Invalid stochastic constraint type!");
        }
      }
      else {
        INTERMEDIATE_econ_[i] = INPUT_econ_[i];
        INTERMEDIATE_emul_[i] = INPUT_emul_[i];
      }
    }
    // Set vector and bound constraint
    buildRiskVec(INPUT_sol_);
    buildRiskBnd(INPUT_bnd_);

    isInitialized_ = false;
  }

  void setStochasticEquality(ParameterList &parlist,
                             const Ptr<SampleGenerator<Real>> &xsampler,
                             const Ptr<BatchManager<Real>> &cbman) {
    std::vector<ParameterList> cparlist(1,parlist);
    std::vector<Ptr<SampleGenerator<Real>>> cxsampler(1,xsampler);
    std::vector<Ptr<BatchManager<Real>>> ccbman(1,cbman);
    setStochasticEquality(cparlist,cxsampler,ccbman);
  }

  /* Inequality constraint */
  void setMeanValueInequality(const Ptr<SampleGenerator<Real>> &sampler,
                              const int index = 0) {
    initStochastic();
    ixsampler_[index] = sampler;
    if ( INPUT_icon_[index] != nullPtr && sampler != nullPtr ) {
      std::vector<Real> mean = computeSampleMean(sampler);
      INTERMEDIATE_icon_[index] = INPUT_icon_[index];
      INTERMEDIATE_icon_[index]->setParameter(mean);
      INTERMEDIATE_ibnd_[index] = INPUT_ibnd_[index];
      INTERMEDIATE_imul_[index] = INPUT_imul_[index];
    }
    else {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::SetMeanValueInequality: Either Constraint or SampleGenerator is NULL!");
    }
    // Set vector and bound constraint
    buildRiskVec(INPUT_sol_);
    buildRiskBnd(INPUT_bnd_);

    isInitialized_ = false;
  }

  void setRiskNeutralInequality(const Ptr<SampleGenerator<Real>> &xsampler,
                                const Ptr<BatchManager<Real>>    &cbman,
                                const int index = 0) {
    initStochastic();
    ixsampler_[index] = xsampler;
    icbman_[index]    = cbman;
    if ( INPUT_icon_[index] != nullPtr
         &&    xsampler     != nullPtr
         &&       cbman     != nullPtr ) {
      INTERMEDIATE_icon_[index]
        = makePtr<RiskNeutralConstraint<Real>>(INPUT_icon_[index],xsampler,cbman);
      INTERMEDIATE_ibnd_[index] = INPUT_ibnd_[index];
      INTERMEDIATE_imul_[index] = INPUT_imul_[index];
    }
    else {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::SetRiskNeutralInequality: Either Constraint, SampleGenerator or BatchManager is NULL!");
    }
    // Set vector and bound constraint
    buildRiskVec(INPUT_sol_);
    buildRiskBnd(INPUT_bnd_);

    isInitialized_ = false;
  }

  void setRiskAverseInequality(ParameterList &parlist,
                               const Ptr<SampleGenerator<Real>> &sampler,
                               const int index = 0) {
    initStochastic();
    ixsampler_[index] = sampler;
    if ( INPUT_icon_[index] != nullPtr && sampler != nullPtr ) {
      needRiskLessIcon_[index] = false;
      parlistCon_[index]       = makePtrFromRef(parlist);
      INTERMEDIATE_icon_[index]
        = makePtr<StochasticConstraint<Real>>(INPUT_icon_[index],sampler,parlist,index);
      INTERMEDIATE_ibnd_[index] = INPUT_ibnd_[index];
      INTERMEDIATE_imul_[index] = INPUT_imul_[index];
    }
    else {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::SetRiskAverseInequality: Either Constraint or SampleGenerator is NULL!");
    }
    // Set vector and bound constraint
    buildRiskVec(INPUT_sol_);
    buildRiskBnd(INPUT_bnd_);

    isInitialized_ = false;
  }

  void setAlmostSureInequality(const Ptr<SampleGenerator<Real>> &sampler,
                               const int index = 0) {
    initStochastic();
    ixsampler_[index] = sampler;
    if ( INPUT_icon_[index] != nullPtr && sampler != nullPtr ) {
      int nsamp = sampler->numMySamples();
      INTERMEDIATE_icon_[index]
        = makePtr<AlmostSureConstraint<Real>>(sampler, INPUT_icon_[index]);
      std::vector<Ptr<Vector<Real>>> imul(nsamp,nullPtr);
      for (int j = 0; j < nsamp; ++j) {
        imul[j] = INPUT_imul_[index]->clone();
        imul[j]->set(*INPUT_imul_[index]);
      }
      INTERMEDIATE_imul_[index]
        = makePtr<DualSimulatedVector<Real>>(imul, sampler->getBatchManager(), sampler);
      INTERMEDIATE_ibnd_[index]
        = makePtr<SimulatedBoundConstraint<Real>>(sampler, INPUT_ibnd_[index]);
    }
    else {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::SetAlmostSureInequality: Either Constraint or SampleGenerator is NULL!");
    }
    // Set vector and bound constraint
    buildRiskVec(INPUT_sol_);
    buildRiskBnd(INPUT_bnd_);

    isInitialized_ = false;
  }

  void setStochasticInequality(std::vector<ParameterList> &parlist,
                               const std::vector<Ptr<SampleGenerator<Real>>> &xsampler,
                               const std::vector<Ptr<BatchManager<Real>>>    &cbman) {
    initStochastic();
    int nc = static_cast<int>(INPUT_icon_.size());
    if ( nc != static_cast<int>(xsampler.size()) || nc != static_cast<int>(cbman.size()) ) {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::setStochasticInequality: Constraint vector and SampleGenerator vector are not the same size!");
    }
    for (int i = 0; i < nc; ++i) {
      if ( xsampler[i] != nullPtr ) {
        std::string type = parlist[i].sublist("SOL").get("Stochastic Component Type","Risk Neutral");
        if ( type == "Risk Neutral" ) {
          setRiskNeutralInequality(xsampler[i],cbman[i],i);
        }
        else if ( type == "Risk Averse" ||
                  type == "Deviation"   ||
                  type == "Error"       ||
                  type == "Regret"      ||
                  type == "Probability" ) {
          setRiskAverseInequality(parlist[i],xsampler[i],i);
        }
        else if ( type == "Almost Sure" ) {
          setAlmostSureInequality(xsampler[i],i);
        }
        else if ( type == "Mean Value" ) {
          setMeanValueInequality(xsampler[i],i);
        }
        else {
          throw Exception::NotImplemented(">>> ROL::OptimizationProblem::SetStochasticInequality: Invalid stochastic constraint type!");
        }
      }
      else {
        INTERMEDIATE_icon_[i] = INPUT_icon_[i];
        INTERMEDIATE_imul_[i] = INPUT_imul_[i];
        INTERMEDIATE_ibnd_[i] = INPUT_ibnd_[i];
      }
    }
    // Set vector and bound constraint
    buildRiskVec(INPUT_sol_);
    buildRiskBnd(INPUT_bnd_);

    isInitialized_ = false;
  }

  void setStochasticInequality(ParameterList &parlist,
                               const Ptr<SampleGenerator<Real>> &xsampler,
                               const Ptr<BatchManager<Real>> &cbman) {
    std::vector<ParameterList> cparlist(1,parlist);
    std::vector<Ptr<SampleGenerator<Real>>> cxsampler(1,xsampler);
    std::vector<Ptr<BatchManager<Real>>> ccbman(1,cbman);
    setStochasticInequality(cparlist,cxsampler,ccbman);
  }

  Real getSolutionStatistic(int comp = 0, int index = 0) {
    Real val(0);
    if (comp == 0) {
      try {
        val = dynamicPtrCast<StochasticObjective<Real>>(INTERMEDIATE_obj_)->computeStatistic(*INTERMEDIATE_sol_);
      }
      catch (std::exception &e) {
        throw Exception::NotImplemented(">>> ROL::OptimizationProblem::getSolutionStatistic: Objective does not have computeStatistic function!");
      }
    }
    else if (comp == 1) {
      int np = INTERMEDIATE_icon_.size();
      if (np <= index || index < 0) {
        throw Exception::NotImplemented(">>> ROL::OptimizationProblem::getSolutionStatistic: Index out of bounds!");
      }
      try {
        val = dynamicPtrCast<StochasticConstraint<Real>>(INTERMEDIATE_icon_[index])->computeStatistic(*INTERMEDIATE_sol_);
      }
      catch (std::exception &e) {
        throw Exception::NotImplemented(">>> ROL::OptimizationProblem::getSolutionStatistic: Constraint does not have computeStatistic function!");
      }
    }
    else {
      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::getSolutionStatistic: Component must be either 0 or 1!");
    }
    return val;
  }

  std::vector<Real> getObjectiveStatistic(void) const {
    try {
      Ptr<std::vector<Real>> stat
        = dynamicPtrCast<RiskVector<Real>>(INTERMEDIATE_sol_)->getStatistic();
      if (stat != nullPtr) {
        return *stat;
      }
      else {
        std::vector<Real> empty;
        return empty;
      }
    }
    catch (std::exception &e) {
      std::vector<Real> empty;
      return empty;
    }
  }

  std::vector<Real> getConstraintStatistic(const int index = 0) const {
    try {
      Ptr<std::vector<Real>> stat
        = dynamicPtrCast<RiskVector<Real>>(INTERMEDIATE_sol_)->getStatistic(1,index);
      if (stat != nullPtr) {
        return *stat;
      }
      else {
        std::vector<Real> empty;
        return empty;
      }
    }
    catch (std::exception &e) {
      std::vector<Real> empty;
      return empty;
    }
  }

  void reset(void) {
    initialize(INTERMEDIATE_obj_,INTERMEDIATE_sol_,INTERMEDIATE_bnd_,
               INTERMEDIATE_econ_,INTERMEDIATE_emul_,
               INTERMEDIATE_icon_,INTERMEDIATE_imul_,INTERMEDIATE_ibnd_);
    conManager_->resetSlackVariables();
  }

  // Check derivatives, and consistency 
  void checkSolutionVector( Vector<Real> &x, // Optimization space
                            Vector<Real> &y, // Optimization space
                            Vector<Real> &u, // Optimization space
                            std::ostream &outStream = std::cout ) {
    initialize(INTERMEDIATE_obj_,INTERMEDIATE_sol_,INTERMEDIATE_bnd_,
               INTERMEDIATE_econ_,INTERMEDIATE_emul_,
               INTERMEDIATE_icon_,INTERMEDIATE_imul_,INTERMEDIATE_ibnd_);
    if (obj_ != nullPtr) {
      outStream << "\nPerforming OptimizationProblem diagnostics." << std::endl << std::endl;

      outStream << "Checking vector operations in optimization vector space X." << std::endl;
      x.checkVector(y,u,true,outStream);
    }
  }

  void checkObjective( Vector<Real> &x, // Optimization space
                       Vector<Real> &u, // Optimization space
                       Vector<Real> &v, // Optimization space
                       std::ostream &outStream = std::cout,
                       const int numSteps = ROL_NUM_CHECKDERIV_STEPS,
                       const int order = 1 ) {
    initialize(INTERMEDIATE_obj_,INTERMEDIATE_sol_,INTERMEDIATE_bnd_,
               INTERMEDIATE_econ_,INTERMEDIATE_emul_,
               INTERMEDIATE_icon_,INTERMEDIATE_imul_,INTERMEDIATE_ibnd_);
    if (obj_ != nullPtr) {
      outStream << "\nPerforming OptimizationProblem diagnostics." << std::endl << std::endl;

      outStream << "Checking objective function." << std::endl;
      obj_->checkGradient(x,v,true,outStream,numSteps,order); outStream << std::endl;
      obj_->checkHessVec(x,u,true,outStream,numSteps,order);  outStream << std::endl;
      obj_->checkHessSym(x,u,v,true,outStream);               outStream << std::endl;
    }
  }

  void checkMultiplierVector( Vector<Real> &w, // Dual constraint space
                              Vector<Real> &q, // Dual constraint space
                              Vector<Real> &l, // Dual constraint space
                              std::ostream &outStream = std::cout ) {
    initialize(INTERMEDIATE_obj_,INTERMEDIATE_sol_,INTERMEDIATE_bnd_,
               INTERMEDIATE_econ_,INTERMEDIATE_emul_,
               INTERMEDIATE_icon_,INTERMEDIATE_imul_,INTERMEDIATE_ibnd_);
    if(con_ != nullPtr) {
      outStream << "\nPerforming OptimizationProblem diagnostics." << std::endl << std::endl;

      outStream << "Checking vector operations in constraint multiplier space C*." << std::endl;
      l.checkVector(q,w,true,outStream);
    }
  }

  void checkConstraint( Vector<Real> &x, // Optimization space
                        Vector<Real> &u, // Optimization space
                        Vector<Real> &v, // Optimization space
                        Vector<Real> &c, // Constraint space
                        Vector<Real> &l, // Dual constraint space
                        std::ostream &outStream = std::cout,
                        const int numSteps = ROL_NUM_CHECKDERIV_STEPS,
                        const int order = 1 ) {
    initialize(INTERMEDIATE_obj_,INTERMEDIATE_sol_,INTERMEDIATE_bnd_,
               INTERMEDIATE_econ_,INTERMEDIATE_emul_,
               INTERMEDIATE_icon_,INTERMEDIATE_imul_,INTERMEDIATE_ibnd_);
    if(con_ != nullPtr) {
      outStream << "\nPerforming OptimizationProblem diagnostics." << std::endl << std::endl;

      outStream << "Checking equality constraint." << std::endl;
      con_->checkApplyJacobian(x,v,c,true,outStream,numSteps,order);         outStream << std::endl;
      con_->checkAdjointConsistencyJacobian(l,u,x,true,outStream);           outStream << std::endl;
      con_->checkApplyAdjointHessian(x,l,v,u,true,outStream,numSteps,order); outStream << std::endl;  
    }
  }

  // Check derivatives, and consistency 
  void check( std::ostream &outStream = std::cout,
              const int numSteps = ROL_NUM_CHECKDERIV_STEPS,
              const int order = 1 ) {
    initialize(INTERMEDIATE_obj_,INTERMEDIATE_sol_,INTERMEDIATE_bnd_,
               INTERMEDIATE_econ_,INTERMEDIATE_emul_,
               INTERMEDIATE_icon_,INTERMEDIATE_imul_,INTERMEDIATE_ibnd_);

    Ptr<Vector<Real>> x, y, u, v;
    try {
      x = sol_->clone(); x->randomize();
      y = sol_->clone(); y->randomize();
      u = sol_->clone(); u->randomize();
      v = sol_->clone(); v->randomize();

      checkSolutionVector(*x,*y,*u,outStream);
      checkObjective(*x,*u,*v,outStream,numSteps,order);
    }
    catch (std::exception &e) {
//      throw Exception::NotImplemented(">>> ROL::OptimizationProblem::check: Elementwise is not implemented for optimization space vectors");
    }

    if(con_ != nullPtr) {
      Ptr<Vector<Real>> c, l, w, q;
      try {
        c = mul_->dual().clone(); c->randomize();
        l = mul_->clone();        l->randomize();
        w = mul_->clone();        w->randomize();
        q = mul_->clone();        q->randomize();

        checkMultiplierVector(*w,*q,*l,outStream);
        checkConstraint(*x,*u,*v,*c,*l,outStream,numSteps,order);
      }
      catch (std::exception &e) {
        throw Exception::NotImplemented(">>> ROL::OptimizationProblem::check: Elementwise is not implemented for constraint space vectors");
      }
    }
  }

}; // class OptimizationProblem

}  // namespace ROL

#endif // ROL_OPTIMIZATIONPROBLEM_HPP