edu.stanford.rsl.jpop

Class ParallelFunctionController

java.lang.Object

edu.stanford.rsl.jpop.FunctionController

edu.stanford.rsl.jpop.ParallelFunctionController

public class ParallelFunctionController
extends FunctionController

Class to process the gradient evaluations in parallel to speed up computationally expensive function / gradient / Hessian evaluations.

Constructor Summary

Constructors

Constructor and Description
ParallelFunctionController()

Method Summary

Modifier and Type	Method and Description
double	evaluate(OptimizableFunction function, double[] x) Evaluates the given Function and returns the result.
double[]	gradient(GradientOptimizableFunction function, double[] x) Evaluates the given Function's Gradient and returns the result.
double[][]	hessian(HessianOptimizableFunction function, double[] x) Evaluates the given Function's Hessian and returns the result.

Methods inherited from class edu.stanford.rsl.jpop.FunctionController

getAssembler, setAssembler

Methods inherited from class java.lang.Object

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

Constructor Detail

ParallelFunctionController

public ParallelFunctionController()

Method Detail

evaluate

public double evaluate(OptimizableFunction function,
                       double[] x)

Description copied from class: FunctionController

Evaluates the given Function and returns the result.

Specified by:

evaluate in class FunctionController

Returns:

gradient

public double[] gradient(GradientOptimizableFunction function,
                         double[] x)

Description copied from class: FunctionController

Evaluates the given Function's Gradient and returns the result.

Specified by:

gradient in class FunctionController

Returns:

hessian

public double[][] hessian(HessianOptimizableFunction function,
                          double[] x)

Description copied from class: FunctionController

Evaluates the given Function's Hessian and returns the result.

Specified by:

hessian in class FunctionController

Returns: