costfunction.hpp

Go to the documentation of this file.

/* -*- mode: c++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
 
/*
 Copyright (C) 2001, 2002, 2003 Nicolas Di Césaré
 Copyright (C) 2015 Peter Caspers
 
 This file is part of QuantLib, a free-software/open-source library
 for financial quantitative analysts and developers - http://quantlib.org/
 
 QuantLib is free software: you can redistribute it and/or modify it
 under the terms of the QuantLib license.  You should have received a
 copy of the license along with this program; if not, please email
 <quantlib-dev@lists.sf.net>. The license is also available online at
 <http://quantlib.org/license.shtml>.
 
 This program 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 license for more details.
*/
 
/*! \file costfunction.hpp
    \brief Optimization cost function class
*/
 
#ifndef quantlib_optimization_costfunction_h
#define quantlib_optimization_costfunction_h
 
#include <ql/math/array.hpp>
#include <ql/math/matrix.hpp>
 
namespace QuantLib {
 
    //!  Cost function abstract class for optimization problem
    class CostFunction {
      public:
        virtual ~CostFunction() = default;
        //! method to overload to compute the cost function value in x
        virtual Real value(const Array& x) const {
            Array v = values(x);
            std::transform(v.begin(), v.end(), v.begin(), [](Real x) -> Real { return x*x; });
            return std::sqrt(std::accumulate(v.begin(), v.end(), Real(0.0)) /
                             static_cast<Real>(v.size()));
        }
        //! method to overload to compute the cost function values in x
        virtual Array values(const Array& x) const =0;
 
        //! method to overload to compute grad_f, the first derivative of
        //  the cost function with respect to x
        virtual void gradient(Array& grad, const Array& x) const {
            Real eps = finiteDifferenceEpsilon(), fp, fm;
            Array xx(x);
            for (Size i=0; i<x.size(); i++) {
                xx[i] += eps;
                fp = value(xx);
                xx[i] -= 2.0*eps;
                fm = value(xx);
                grad[i] = 0.5*(fp - fm)/eps;
                xx[i] = x[i];
            }
        }
 
        //! method to overload to compute grad_f, the first derivative of
        //  the cost function with respect to x and also the cost function
        virtual Real valueAndGradient(Array& grad,
                                      const Array& x) const {
            gradient(grad, x);
            return value(x);
        }
 
        //! method to overload to compute J_f, the jacobian of
        // the cost function with respect to x
        virtual void jacobian(Matrix &jac, const Array &x) const {
            Real eps = finiteDifferenceEpsilon();
            Array xx(x), fp, fm;
            for(Size i=0; i<x.size(); ++i) {
                xx[i] += eps;
                fp = values(xx);
                xx[i] -= 2.0*eps;
                fm = values(xx);
                for(Size j=0; j<fp.size(); ++j) {
                    jac[j][i] = 0.5*(fp[j]-fm[j])/eps;
                }
                xx[i] = x[i];
            }
        }
 
        //! method to overload to compute J_f, the jacobian of
        // the cost function with respect to x and also the cost function
        virtual Array valuesAndJacobian(Matrix &jac,
                                        const Array &x) const {
            jacobian(jac,x);
            return values(x);
        }
 
        //! Default epsilon for finite difference method :
        virtual Real finiteDifferenceEpsilon() const { return 1e-8; }
    };
 
    class ParametersTransformation {
      public:
        virtual ~ParametersTransformation() = default;
        virtual Array direct(const Array& x) const = 0;
        virtual Array inverse(const Array& x) const = 0;
    };
}
 
#endif