edu.stanford.nlp.optimization
Class PenaltyConstrainedMinimizer

java.lang.Object
  |
  +--edu.stanford.nlp.optimization.PenaltyConstrainedMinimizer

All Implemented Interfaces:

ConstrainedMinimizer, Minimizer

public class PenaltyConstrainedMinimizer

extends Object

implements ConstrainedMinimizer

A wrapper class which builds a constrained minimizer out of an unconstrained one by finding the unconstrained minimum of a sequence of penalized surfaces. It assumes that the objective it is given is a DiffFunction, even if the base unconstrained minimzer just ignores the derivative information. The desired unconstrained minimizer is passed in on construction: Minimizer m = new SomeUnconstrainedMinimizer(); ConstrainedMinimizer cm = new PenaltyConstrainedMinimizer(m);

Since:

1.0

See Also:

ConstrainedMinimizer

Constructor Summary

PenaltyConstrainedMinimizer(Minimizer minimizer)

Method Summary

double[]	minimize(Function function, double functionTolerance, double[] initial) Attempts to find an unconstrained minimum of the objective function starting at initial, within functionTolerance.
double[]	minimize(Function function, double functionTolerance, Function[] eqConstraints, double eqConstraintTolerance, Function[] ineqConstraints, double ineqConstraintTolerance, double[] initial) The general case, allowing both equality and inequality constraints.

Methods inherited from class java.lang.Object

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

Constructor Detail

PenaltyConstrainedMinimizer

public PenaltyConstrainedMinimizer(Minimizer minimizer)

Method Detail

minimize

public double[] minimize(Function function,
                         double functionTolerance,
                         Function[] eqConstraints,
                         double eqConstraintTolerance,
                         Function[] ineqConstraints,
                         double ineqConstraintTolerance,
                         double[] initial)

Description copied from interface: ConstrainedMinimizer

The general case, allowing both equality and inequality constraints. A given implementation can throw an UnsupportedOperationException if it cannot handle the constraints it is supplied. It is expected that minimize(function, functionTolerance, [], 0, [], 0, initial) behave exactly like the minimize(function, functionTolerance, initial) method inherited from Minimizer.

Specified by:

minimize in interface ConstrainedMinimizer

Parameters:

function - the objective function

functionTolerance - a double value

eqConstraints - an array of zero or more equality constraints

eqConstraintTolerance - the violation tolerace for equality constraints

ineqConstraints - an array of zero or more inequality constraints

ineqConstraintTolerance - the violation tolerace for equality constraints

initial - a initial feasible (!) point

Returns:

a minimizing feasible point

minimize

public double[] minimize(Function function,
                         double functionTolerance,
                         double[] initial)

Description copied from interface: Minimizer

Attempts to find an unconstrained minimum of the objective function starting at initial, within functionTolerance.

Specified by:

minimize in interface Minimizer

Parameters:

function - the objective function

functionTolerance - a double value

initial - a initial feasible point