/// <summary>
/// Non-linear least square fitting by the trust-region algorithm.
/// </summary>
/// <param name="objective">The objective model, including function, jacobian, observations, and parameter bounds.</param>
/// <param name="subproblem">The subproblem</param>
/// <param name="initialGuess">The initial guess values.</param>
/// <param name="functionTolerance">The stopping threshold for L2 norm of the residuals.</param>
/// <param name="gradientTolerance">The stopping threshold for infinity norm of the gradient vector.</param>
/// <param name="stepTolerance">The stopping threshold for L2 norm of the change of parameters.</param>
/// <param name="radiusTolerance">The stopping threshold for trust region radius</param>
/// <param name="maximumIterations">The max iterations.</param>
/// <returns></returns>
public NonlinearMinimizationResult Minimum(ITrustRegionSubproblem subproblem, IObjectiveModel objective, Vector <double> initialGuess,
Vector <double> lowerBound = null, Vector <double> upperBound = null, Vector <double> scales = null, List <bool> isFixed = null,
double gradientTolerance = 1E-8, double stepTolerance = 1E-8, double functionTolerance = 1E-8, double radiusTolerance = 1E-18, int maximumIterations = -1)
{
// Non-linear least square fitting by the trust-region algorithm.
//
// For given datum pair (x, y), uncertainties σ (or weighting W = 1 / σ^2) and model function f = f(x; p),
// let's find the parameters of the model so that the sum of the quares of the deviations is minimized.
//
// F(p) = 1/2 * ∑{ Wi * (yi - f(xi; p))^2 }
// pbest = argmin F(p)
//
// Here, we will use the following terms:
// Weighting W is the diagonal matrix and can be decomposed as LL', so L = 1/σ
// Residuals, R = L(y - f(x; p))
// Residual sum of squares, RSS = ||R||^2 = R.DotProduct(R)
// Jacobian J = df(x; p)/dp
// Gradient g = -J'W(y − f(x; p)) = -J'LR
// Approximated Hessian H = J'WJ
//
// The trust region algorithm is summarized as follows:
// initially set trust-region radius, Δ
// repeat
// solve subproblem
// update Δ:
// let ρ = (RSS - RSSnew) / predRed
// if ρ > 0.75, Δ = 2Δ
// if ρ < 0.25, Δ = Δ/4
// if ρ > eta, P = P + ΔP
//
// References:
// [1]. Madsen, K., H. B. Nielsen, and O. Tingleff.
// "Methods for Non-Linear Least Squares Problems. Technical University of Denmark, 2004. Lecture notes." (2004).
// Available Online from: http://orbit.dtu.dk/files/2721358/imm3215.pdf
// [2]. Nocedal, Jorge, and Stephen J. Wright.
// Numerical optimization (2006): 101-134.
// [3]. SciPy
// Available Online from: https://github.com/scipy/scipy/blob/master/scipy/optimize/_trustregion.py
double maxDelta = 1000;
double eta = 0;
if (objective == null)
{
throw new ArgumentNullException(nameof(objective));
}
ValidateBounds(initialGuess, lowerBound, upperBound, scales);
objective.SetParameters(initialGuess, isFixed);
ExitCondition exitCondition = ExitCondition.None;
// First, calculate function values and setup variables
var P = ProjectToInternalParameters(initialGuess); // current internal parameters
Vector <double> Pstep; // the change of parameters
var RSS = EvaluateFunction(objective, initialGuess); // Residual Sum of Squares
if (maximumIterations < 0)
{
maximumIterations = 200 * (initialGuess.Count + 1);
}
// if RSS == NaN, stop
if (double.IsNaN(RSS))
{
exitCondition = ExitCondition.InvalidValues;
return(new NonlinearMinimizationResult(objective, -1, exitCondition));
}
// When only function evaluation is needed, set maximumIterations to zero,
if (maximumIterations == 0)
{
exitCondition = ExitCondition.ManuallyStopped;
}
// if ||R||^2 <= fTol, stop
if (RSS <= functionTolerance)
{
exitCondition = ExitCondition.Converged; // SmallRSS
}
// evaluate projected gradient and Hessian
var(Gradient, Hessian) = EvaluateJacobian(objective, P);
// if ||g||_oo <= gtol, found and stop
if (Gradient.InfinityNorm() <= gradientTolerance)
{
exitCondition = ExitCondition.RelativeGradient; // SmallGradient
}
if (exitCondition != ExitCondition.None)
{
return(new NonlinearMinimizationResult(objective, -1, exitCondition));
}
// initialize trust-region radius, Δ
double delta = Gradient.DotProduct(Gradient) / (Hessian * Gradient).DotProduct(Gradient);
delta = Math.Max(1.0, Math.Min(delta, maxDelta));
int iterations = 0;
bool hitBoundary;
while (iterations < maximumIterations && exitCondition == ExitCondition.None)
{
iterations++;
// solve the subproblem
subproblem.Solve(objective, delta);
Pstep = subproblem.Pstep;
hitBoundary = subproblem.HitBoundary;
// predicted reduction = L(0) - L(Δp) = -Δp'g - 1/2 * Δp'HΔp
var predictedReduction = -Gradient.DotProduct(Pstep) - 0.5 * Pstep.DotProduct(Hessian * Pstep);
if (Pstep.L2Norm() <= stepTolerance * (stepTolerance + P.L2Norm()))
{
exitCondition = ExitCondition.RelativePoints; // SmallRelativeParameters
break;
}
var Pnew = P + Pstep; // parameters to test
// evaluate function at Pnew
var RSSnew = EvaluateFunction(objective, Pnew);
// if RSS == NaN, stop
if (double.IsNaN(RSSnew))
{
exitCondition = ExitCondition.InvalidValues;
break;
}
// calculate the ratio of the actual to the predicted reduction.
double rho = (predictedReduction != 0)
? (RSS - RSSnew) / predictedReduction
: 0.0;
if (rho > 0.75 && hitBoundary)
{
delta = Math.Min(2.0 * delta, maxDelta);
}
else if (rho < 0.25)
{
delta = delta * 0.25;
if (delta <= radiusTolerance * (radiusTolerance + P.DotProduct(P)))
{
exitCondition = ExitCondition.LackOfProgress;
break;
}
}
if (rho > eta)
{
// accepted
Pnew.CopyTo(P);
RSS = RSSnew;
// evaluate projected gradient and Hessian
(Gradient, Hessian) = EvaluateJacobian(objective, P);
// if ||g||_oo <= gtol, found and stop
if (Gradient.InfinityNorm() <= gradientTolerance)
{
exitCondition = ExitCondition.RelativeGradient;
}
// if ||R||^2 < fTol, found and stop
if (RSS <= functionTolerance)
{
exitCondition = ExitCondition.Converged; // SmallRSS
}
}
}
if (iterations >= maximumIterations)
{
exitCondition = ExitCondition.ExceedIterations;
}
return(new NonlinearMinimizationResult(objective, iterations, exitCondition));
}
/// <summary>
/// Non-linear least square fitting by the Levenberg-Marduardt algorithm.
/// </summary>
/// <param name="objective">The objective function, including model, observations, and parameter bounds.</param>
/// <param name="initialGuess">The initial guess values.</param>
/// <param name="initialMu">The initial damping parameter of mu.</param>
/// <param name="gradientTolerance">The stopping threshold for infinity norm of the gradient vector.</param>
/// <param name="stepTolerance">The stopping threshold for L2 norm of the change of parameters.</param>
/// <param name="functionTolerance">The stopping threshold for L2 norm of the residuals.</param>
/// <param name="maximumIterations">The max iterations.</param>
/// <returns>The result of the Levenberg-Marquardt minimization</returns>
public NonlinearMinimizationResult Minimum(IObjectiveModel objective, Vector <double> initialGuess,
Vector <double> lowerBound = null, Vector <double> upperBound = null, Vector <double> scales = null, List <bool> isFixed = null,
double initialMu = 1E-3, double gradientTolerance = 1E-15, double stepTolerance = 1E-15, double functionTolerance = 1E-15, int maximumIterations = -1)
{
// Non-linear least square fitting by the Levenberg-Marduardt algorithm.
//
// Levenberg-Marquardt is finding the minimum of a function F(p) that is a sum of squares of nonlinear functions.
//
// For given datum pair (x, y), uncertainties σ (or weighting W = 1 / σ^2) and model function f = f(x; p),
// let's find the parameters of the model so that the sum of the quares of the deviations is minimized.
//
// F(p) = 1/2 * ∑{ Wi * (yi - f(xi; p))^2 }
// pbest = argmin F(p)
//
// We will use the following terms:
// Weighting W is the diagonal matrix and can be decomposed as LL', so L = 1/σ
// Residuals, R = L(y - f(x; p))
// Residual sum of squares, RSS = ||R||^2 = R.DotProduct(R)
// Jacobian J = df(x; p)/dp
// Gradient g = -J'W(y − f(x; p)) = -J'LR
// Approximated Hessian H = J'WJ
//
// The Levenberg-Marquardt algorithm is summarized as follows:
// initially let μ = τ * max(diag(H)).
// repeat
// solve linear equations: (H + μI)ΔP = -g
// let ρ = (||R||^2 - ||Rnew||^2) / (Δp'(μΔp - g)).
// if ρ > ε, P = P + ΔP; μ = μ * max(1/3, 1 - (2ρ - 1)^3); ν = 2;
// otherwise μ = μ*ν; ν = 2*ν;
//
// References:
// [1]. Madsen, K., H. B. Nielsen, and O. Tingleff.
// "Methods for Non-Linear Least Squares Problems. Technical University of Denmark, 2004. Lecture notes." (2004).
// Available Online from: http://orbit.dtu.dk/files/2721358/imm3215.pdf
// [2]. Gavin, Henri.
// "The Levenberg-Marquardt method for nonlinear least squares curve-fitting problems."
// Department of Civil and Environmental Engineering, Duke University (2017): 1-19.
// Availble Online from: http://people.duke.edu/~hpgavin/ce281/lm.pdf
if (objective == null)
{
throw new ArgumentNullException(nameof(objective));
}
ValidateBounds(initialGuess, lowerBound, upperBound, scales);
objective.SetParameters(initialGuess, isFixed);
ExitCondition exitCondition = ExitCondition.None;
// First, calculate function values and setup variables
var P = ProjectToInternalParameters(initialGuess); // current internal parameters
Vector <double> Pstep; // the change of parameters
var RSS = EvaluateFunction(objective, P); // Residual Sum of Squares = R'R
if (maximumIterations < 0)
{
maximumIterations = 200 * (initialGuess.Count + 1);
}
// if RSS == NaN, stop
if (double.IsNaN(RSS))
{
exitCondition = ExitCondition.InvalidValues;
return(new NonlinearMinimizationResult(objective, -1, exitCondition));
}
// When only function evaluation is needed, set maximumIterations to zero,
if (maximumIterations == 0)
{
exitCondition = ExitCondition.ManuallyStopped;
}
// if RSS <= fTol, stop
if (RSS <= functionTolerance)
{
exitCondition = ExitCondition.Converged; // SmallRSS
}
// Evaluate gradient and Hessian
var(Gradient, Hessian) = EvaluateJacobian(objective, P);
var diagonalOfHessian = Hessian.Diagonal(); // diag(H)
// if ||g||oo <= gtol, found and stop
if (Gradient.InfinityNorm() <= gradientTolerance)
{
exitCondition = ExitCondition.RelativeGradient;
}
if (exitCondition != ExitCondition.None)
{
return(new NonlinearMinimizationResult(objective, -1, exitCondition));
}
double mu = initialMu * diagonalOfHessian.Max(); // μ
double nu = 2; // ν
int iterations = 0;
while (iterations < maximumIterations && exitCondition == ExitCondition.None)
{
iterations++;
while (true)
{
Hessian.SetDiagonal(Hessian.Diagonal() + mu); // hessian[i, i] = hessian[i, i] + mu;
// solve normal equations
Pstep = Hessian.Solve(-Gradient);
// if ||ΔP|| <= xTol * (||P|| + xTol), found and stop
if (Pstep.L2Norm() <= stepTolerance * (stepTolerance + P.DotProduct(P)))
{
exitCondition = ExitCondition.RelativePoints;
break;
}
var Pnew = P + Pstep; // new parameters to test
// evaluate function at Pnew
var RSSnew = EvaluateFunction(objective, Pnew);
if (double.IsNaN(RSSnew))
{
exitCondition = ExitCondition.InvalidValues;
break;
}
// calculate the ratio of the actual to the predicted reduction.
// ρ = (RSS - RSSnew) / (Δp'(μΔp - g))
var predictedReduction = Pstep.DotProduct(mu * Pstep - Gradient);
var rho = (predictedReduction != 0)
? (RSS - RSSnew) / predictedReduction
: 0;
if (rho > 0.0)
{
// accepted
Pnew.CopyTo(P);
RSS = RSSnew;
// update gradient and Hessian
(Gradient, Hessian) = EvaluateJacobian(objective, P);
diagonalOfHessian = Hessian.Diagonal();
// if ||g||_oo <= gtol, found and stop
if (Gradient.InfinityNorm() <= gradientTolerance)
{
exitCondition = ExitCondition.RelativeGradient;
}
// if ||R||^2 < fTol, found and stop
if (RSS <= functionTolerance)
{
exitCondition = ExitCondition.Converged; // SmallRSS
}
mu = mu * Math.Max(1.0 / 3.0, 1.0 - Math.Pow(2.0 * rho - 1.0, 3));
nu = 2;
break;
}
else
{
// rejected, increased μ
mu = mu * nu;
nu = 2 * nu;
Hessian.SetDiagonal(diagonalOfHessian);
}
}
}
if (iterations >= maximumIterations)
{
exitCondition = ExitCondition.ExceedIterations;
}
return(new NonlinearMinimizationResult(objective, iterations, exitCondition));
}
