/// ------------------------------------------------------
/// SwarmOps - Numeric and heuristic optimization for C#
/// Copyright (C) 2003-2011 Magnus Erik Hvass Pedersen.
/// Please see the file license.txt for license details.
/// SwarmOps on the internet: http://www.Hvass-Labs.org/
/// ------------------------------------------------------
using System.Diagnostics;
namespace SwarmOps.Problems
{
///
/// Curve-fitting to the exponential curve f(x) = B*Pow(A, x).
/// You may wish to use a form of regression instead,
/// depending on your requirements for statistical minimization
/// of error.
///
public class CurveFittingExp : CurveFitting
{
#region Constructors.
///
/// Create the object.
///
/// X-axis values.
/// Y-axis values, curve to be fitted.
/// Minimum value for A parameter.
/// Maximum value for A parameter.
/// Minimum value for B parameter.
/// Maximum value for B parameter.
public CurveFittingExp(
double[] x, double[] y,
double minA, double maxA,
double minB, double maxB)
: base(x, y)
{
_lowerBound = new double[] { minA, minB };
_lowerInit = _lowerBound;
_upperBound = new double[] { maxA, maxB };
_upperInit = _upperBound;
}
#endregion
#region Public methods.
///
/// Get the A parameter.
///
public double GetA(double[] parameters)
{
return parameters[0];
}
///
/// Set the A parameter.
///
public void SetA(ref double[] parameters, double a)
{
parameters[0] = a;
}
///
/// Get the B parameter.
///
public double GetB(double[] parameters)
{
return parameters[1];
}
///
/// Set the B parameter.
///
public void SetB(ref double[] parameters, double b)
{
parameters[1] = b;
}
#endregion
#region Base-class overrides, Problem.
///
/// Name of the optimization problem.
///
public override string Name
{
get { return "CurveFittingExp"; }
}
///
/// Dimensionality of the optimization problem.
///
public override int Dimensionality
{
get { return 2; }
}
///
/// Threshold for an acceptable fitness value.
///
public override double AcceptableFitness
{
get
{
return 0.001;
}
}
double[] _lowerBound;
///
/// Lower boundary for the search-space.
///
public override double[] LowerBound
{
get { return _lowerBound; }
}
double[] _upperBound;
///
/// Upper boundary for the search-space.
///
public override double[] UpperBound
{
get { return _upperBound; }
}
double[] _lowerInit;
///
/// Lower initialization boundary.
///
public override double[] LowerInit
{
get { return _lowerInit; }
}
double[] _upperInit;
///
/// Upper initialization boundary.
///
public override double[] UpperInit
{
get { return _upperInit; }
}
string[] _parameterName = { "A", "B" };
///
/// Array with names of parameters.
///
public override string[] ParameterName
{
get { return _parameterName; }
}
///
/// Has the gradient has been implemented?
///
public override bool HasGradient
{
get { return true; }
}
///
/// Compute the gradient of the fitness-function.
///
/// Candidate solution.
/// Array for holding the gradient.
public override int Gradient(double[] parameters, ref double[] v)
{
Debug.Assert(parameters != null && parameters.Length == Dimensionality);
Debug.Assert(v != null && v.Length == Dimensionality);
Debug.Assert(X.Length == Y.Length);
// Get parameters.
double a = GetA(parameters);
double b = GetB(parameters);
double gradientA = 0;
double gradientB = 0;
for (int i = 0; i < X.Length; i++)
{
double x = X[i];
double x2 = x * x;
double lnx = System.Math.Log(x);
double lnx2 = System.Math.Log(x2);
double xPowa = System.Math.Pow(x, a);
double x2Powa = System.Math.Pow(x2, a);
double y = Y[i];
gradientA += b * b * lnx2 * x2Powa - 2 * b * y * lnx * xPowa;
gradientB += 2 * (b * x2Powa - xPowa * y);
}
SetA(ref v, gradientA);
SetB(ref v, gradientB);
return 0;
}
#endregion
#region Base-class overrides, CurveFitting.
///
/// Compute the value y given x using the curve-fitting function.
///
/// Parameters for curve-fitting function.
/// X-axis value.
/// Computed Y-axis value.
public override double ComputeY(double[] parameters, double x)
{
double a = GetA(parameters);
double b = GetB(parameters);
return b * System.Math.Pow(a, x);
}
#endregion
}
}