com.numericalmethod.suanshu.optimization.unconstrained.steepestdescent
Class GaussNewton

java.lang.Object
  com.numericalmethod.suanshu.optimization.unconstrained.steepestdescent.SteepestDescent
      com.numericalmethod.suanshu.optimization.unconstrained.steepestdescent.GaussNewton

All Implemented Interfaces:

Minimizer, UnconstrainedMinimizer

public class GaussNewton
extends SteepestDescent

The Gauss-Newton method is a steepest descent method to minimize a real vector function in the form:

f(x) = [f1(x) f2(x) ... fm(x)]'

The objective function is

F(x) = f' %*% f

See Also:

• "Algorithm 5.5. Practical Optimization: Algorithms and Engineering Applications. Andreas Antoniou and Wu-Sheng Lu."
• Wikipedia: Gauss–Newton algorithm

Nested Class Summary

Nested classes/interfaces inherited from class com.numericalmethod.suanshu.optimization.unconstrained.steepestdescent.SteepestDescent

SteepestDescent.LineSearch

Field Summary

Fields inherited from class com.numericalmethod.suanshu.optimization.unconstrained.steepestdescent.SteepestDescent

f, g, tol

Constructor Summary

GaussNewton()

Method Summary

protected SteepestDescent.LineSearch	getLineSearch()
void	solve(RealVectorFunction vf, double tol) Construct an instance of GaussNewton to minimize a real vector function f.
void	solve(RealVectorFunction vf, RntoMatrix J, double tol) Construct an instance of GaussNewton to minimize a real vector function f.
void	solve(UnconstrainedProblem problem, double tol)

Methods inherited from class com.numericalmethod.suanshu.optimization.unconstrained.steepestdescent.SteepestDescent

minimum, search, solve, solve

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

GaussNewton

public GaussNewton()

Method Detail

getLineSearch

protected SteepestDescent.LineSearch getLineSearch()

Specified by:

getLineSearch in class SteepestDescent

solve

public void solve(UnconstrainedProblem problem,
                  double tol)

Specified by:

solve in interface UnconstrainedMinimizer

Overrides:

solve in class SteepestDescent

solve

public void solve(RealVectorFunction vf,
                  RntoMatrix J,
                  double tol)

Construct an instance of GaussNewton to minimize a real vector function f.

Parameters:

vf - the real vector function to be minimized

J - a function that computes the Jacobian of f for a given x

tol - the convergence tolerance

solve

public void solve(RealVectorFunction vf,
                  double tol)

Construct an instance of GaussNewton to minimize a real vector function f. The Jacobian is computed numerically.

Parameters:

vf - the real vector function to be minimized

tol - the convergence tolerance