edu.stanford.nlp.optimization
Interface Minimizer<T extends Function>

All Known Implementing Classes:

CGMinimizer, HybridMinimizer, QNMinimizer, ScaledSGDMinimizer, SGDMinimizer, SGDToQNMinimizer, SMDMinimizer, SQNMinimizer, StochasticInPlaceMinimizer, StochasticMinimizer

public interface Minimizer<T extends Function>

The interface for unconstrained function minimizers.

Implementations may also vary in their requirements for the arguments. For example, implementations may or may not care if the initial feasible vector turns out to be non-feasible (or null!). Similarly, some methods may insist that objectives and/or constraint Function objects actually be DiffFunction objects.

Since:

1.0

Author:

Dan Klein

Method Summary

double[]	minimize(T function, double functionTolerance, double[] initial) Attempts to find an unconstrained minimum of the objective function starting at initial, within functionTolerance.
double[]	minimize(T function, double functionTolerance, double[] initial, int maxIterations)

Method Detail

minimize

double[] minimize(T function,
                  double functionTolerance,
                  double[] initial)

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

Parameters:

function - the objective function

functionTolerance - a double value

initial - a initial feasible point

Returns:

Unconstrained minimum of function

minimize

double[] minimize(T function,
                  double functionTolerance,
                  double[] initial,
                  int maxIterations)