ROL_BoundConstraint.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_BOUND_CONSTRAINT_H
#define ROL_BOUND_CONSTRAINT_H

#include "ROL_Vector.hpp"
#include "ROL_Types.hpp"
#include <iostream>

namespace ROL {

template <class Real>
class BoundConstraint {
private:
  int dim_;
  Teuchos::RCP<Vector<Real> > x_lo_;
  Teuchos::RCP<Vector<Real> > x_up_;
  Real scale_;  
  bool activated_; 
  Real min_diff_;

  Elementwise::ReductionMin<Real> minimum_;

  // Add a constant to every element in a vector
  class Shift : public Elementwise::UnaryFunction<Real> {
    public:
      Shift(Real offset) : offset_(offset) {}
      Real apply( const Real &x ) const { return offset_ + x; }
    private:
      Real offset_;   
  };
  
  class LogicalOr : public Elementwise::BinaryFunction<Real> {
    public:
      Real apply( const Real &x, const Real &y ) const {
        return (x==0.0 && y==0.0) ? 0.0 : 1.0;
      }
  };

  class Product : public Elementwise::BinaryFunction<Real> {
    public:
      Real apply( const Real &x, const Real &y ) const { return x*y; }
  } product;


public:

  virtual ~BoundConstraint() {}

  BoundConstraint(void) : activated_(true) {}  

  BoundConstraint(const Teuchos::RCP<Vector<Real> > &x_lo,
                  const Teuchos::RCP<Vector<Real> > &x_up,
                  Real scale = 1.0) : 
                    x_lo_(x_lo), x_up_(x_up),scale_(scale),activated_(true) {

    // Compute difference between upper and lower bounds
    Teuchos::RCP<Vector<Real> > diff = x_up_->clone();
    diff->set(*x_up_);
    diff->axpy(-1.0,*x_lo_);  

    // Compute minimum difference 
    min_diff_ = diff->reduce(minimum_);
    min_diff_ *= 0.5; 
    
  }

  virtual void update( const Vector<Real> &x, bool flag = true, int iter = -1 ) {}

  virtual void project( Vector<Real> &x ) {

    struct Lesser : public Elementwise::BinaryFunction<Real> {
      Real apply(const Real &x, const Real &y) const { return x<y ? x : y; }   
    } lesser;
   
    struct Greater : public Elementwise::BinaryFunction<Real> {
      Real apply(const Real &x, const Real &y) const { return x>y ? x : y; }   
    } greater;

    x.applyBinary(lesser, *x_up_); // Set x to the elementwise minimum of x and x_up_
    x.applyBinary(greater,*x_lo_); // Set x to the elementwise minimum of x and x_lo_

  }

  virtual void pruneUpperActive( Vector<Real> &v, const Vector<Real> &x, Real eps = 0.0 ) {

    Real epsn = std::min(scale_*eps,this->min_diff_);          

    // Find the indices where x - u + epsn >= 0

    Shift shift(epsn);    

    Teuchos::RCP<Vector<Real> > mask = x.clone();
    mask->set(x);
    mask->axpy(-1.0,*x_up_);
    mask->applyUnary(shift); 
    
    struct Condition : public Elementwise::UnaryFunction<Real> {
      Real apply(const Real &x) const {
        return x>=0 ? 0.0 : 1.0;
      }
    } condition;   

    mask->applyUnary(condition); // mask = 0 anywhere x - u + epsn >= 0

    v.applyBinary(product, *mask); // zeros elements of v where mask=0 
  }

  virtual void pruneUpperActive( Vector<Real> &v, const Vector<Real> &g, const Vector<Real> &x, Real eps = 0.0 ) {

    Real epsn = std::min(scale_*eps,this->min_diff_);          

    // Find the indices where x - u + epsn >= 0

    Shift shift(epsn);    

    Teuchos::RCP<Vector<Real> > mask1 = x.clone();
    mask1->set(x);
    mask1->axpy(-1.0,*x_up_);
    mask1->applyUnary(shift); 
    
    struct Condition1 : public Elementwise::UnaryFunction<Real> {
      Real apply(const Real &x) const {
        return x>=0 ? 0.0 : 1.0;
      }
    } condition1;   

    mask1->applyUnary(condition1);

    struct Condition2 : public Elementwise::UnaryFunction<Real> {  
      Real apply(const Real &x) const {
        return x < 0.0 ? 0.0 : 1.0;
      }
    } condition2;

    Teuchos::RCP<Vector<Real> > mask2 = g.clone();
    mask2->set(g);    

    mask2->applyUnary(condition2); 

    LogicalOr logicalOr;

    mask1->applyBinary(logicalOr,*mask2); 

    v.applyBinary(product,*mask1);
 
  }
 
  virtual void pruneLowerActive( Vector<Real> &v, const Vector<Real> &x, Real eps = 0.0 ) {

    Real epsn = std::min(scale_*eps,this->min_diff_);          

    // Find the indices where -x + l + epsn >= 0

    Shift shift(epsn);    

    Teuchos::RCP<Vector<Real> > mask = x_lo_->clone();
    mask->set(*x_lo_);
    mask->axpy(-1.0,x);
    mask->applyUnary(shift); 
    
    struct Condition: public Elementwise::UnaryFunction<Real> {
      Real apply(const Real &x) const {
        return x>=0 ? 0.0 : 1.0;
      }
    } condition;   

    mask->applyUnary(condition); // mask = 0 anywhere -x + l + epsn >= 0

    v.applyBinary(product, *mask); // zeros elements of v where mask=0 

  }

  virtual void pruneLowerActive( Vector<Real> &v, const Vector<Real> &g, const Vector<Real> &x, Real eps = 0.0 ) {

    Real epsn = std::min(scale_*eps,this->min_diff_);          

    // Find the indices where -x + l + epsn >= 0

    Shift shift(epsn);    

    Teuchos::RCP<Vector<Real> > mask1 = x_lo_->clone();
    mask1->set(*x_lo_);
    mask1->axpy(-1.0,x);
    mask1->applyUnary(shift); 
    
    struct Condition1: public Elementwise::UnaryFunction<Real> {
      Real apply(const Real &x) const {
        return x>=0 ? 0.0 : 1.0;
      }
    } condition1;   

    mask1->applyUnary(condition1); // mask = 0 anywhere -x + l + epsn >= 0

    struct Condition2 : public Elementwise::UnaryFunction<Real> {  
      Real apply(const Real &x) const {
        return x > 0.0 ? 0.0 : 1.0;
      }
    } condition2;
    
    Teuchos::RCP<Vector<Real> > mask2 = g.clone();
    mask2->set(g); 

    mask2->applyUnary(condition2); 

    LogicalOr logicalOr;

    mask1->applyBinary(logicalOr,*mask2); 

    v.applyBinary(product,*mask1);

  }
 
  virtual void setVectorToUpperBound( Vector<Real> &u ) {
    u.set(*x_up_);
  }

  virtual void setVectorToLowerBound( Vector<Real> &l ) {
    l.set(*x_lo_);
  }

  virtual void pruneActive( Vector<Real> &v, const Vector<Real> &x, Real eps = 0.0 ) {
    this->pruneUpperActive(v,x,eps);
    this->pruneLowerActive(v,x,eps);
  }

  virtual void pruneActive( Vector<Real> &v, const Vector<Real> &g, const Vector<Real> &x, Real eps = 0.0 ) {
    this->pruneUpperActive(v,g,x,eps);
    this->pruneLowerActive(v,g,x,eps);
  }

  virtual bool isFeasible( const Vector<Real> &v ) { 
    if ( this->activated_ ) {
      Teuchos::RCP<Vector<Real> > s = v.clone();
      s->set(*x_up_);

      s->axpy(-1.0,v); // u-v

      Real uminusv = s->reduce(minimum_);

      s->set(v);
      s->axpy(-1.0,*x_lo_);

      Real vminusl = s->reduce(minimum_);

      if( (uminusv<0) || (vminusl<0) ) {
        return false;
      } 
      else {
        return true;
      }

    }
    else {
      return true; 
    }
  }

  void activate(void)    { this->activated_ = true;  }

  void deactivate(void)  { this->activated_ = false; }

  bool isActivated(void) { return this->activated_;  }

  void pruneInactive( Vector<Real> &v, const Vector<Real> &x, Real eps = 0.0 ) { 
    Teuchos::RCP<Vector<Real> > tmp = v.clone(); 
    tmp->set(v);
    this->pruneActive(*tmp,x,eps);
    v.axpy(-1.0,*tmp);
  }
  void pruneLowerInactive( Vector<Real> &v, const Vector<Real> &x, Real eps = 0.0 ) { 
    Teuchos::RCP<Vector<Real> > tmp = v.clone(); 
    tmp->set(v);
    this->pruneLowerActive(*tmp,x,eps);
    v.axpy(-1.0,*tmp);
  }
  void pruneUpperInactive( Vector<Real> &v, const Vector<Real> &x, Real eps = 0.0 ) { 
    Teuchos::RCP<Vector<Real> > tmp = v.clone(); 
    tmp->set(v);
    this->pruneUpperActive(*tmp,x,eps);
    v.axpy(-1.0,*tmp);
  }


  void pruneInactive( Vector<Real> &v, const Vector<Real> &g, const Vector<Real> &x, Real eps = 0.0 ) { 
    Teuchos::RCP<Vector<Real> > tmp = v.clone(); 
    tmp->set(v);
    this->pruneActive(*tmp,g,x,eps);
    v.axpy(-1.0,*tmp);
  }
  void pruneLowerInactive( Vector<Real> &v, const Vector<Real> &g, const Vector<Real> &x, Real eps = 0.0 ) { 
    Teuchos::RCP<Vector<Real> > tmp = v.clone(); 
    tmp->set(v);
    this->pruneLowerActive(*tmp,g,x,eps);
    v.axpy(-1.0,*tmp);
  }
  void pruneUpperInactive( Vector<Real> &v, const Vector<Real> &g, const Vector<Real> &x, Real eps = 0.0 ) { 
    Teuchos::RCP<Vector<Real> > tmp = v.clone(); 
    tmp->set(v);
    this->pruneUpperActive(*tmp,g,x,eps);
    v.axpy(-1.0,*tmp);
  }
 
  void computeProjectedGradient( Vector<Real> &g, const Vector<Real> &x ) {
    Teuchos::RCP<Vector<Real> > tmp = g.clone();
    tmp->set(g);
    this->pruneActive(g,*tmp,x);
  }
 
  void computeProjectedStep( Vector<Real> &v, const Vector<Real> &x ) { 
    v.plus(x);
    this->project(v);
    v.axpy(-1.0,x);
  }

}; // class BoundConstraint

} // namespace ROL

#endif