de.torstennahm.integrate.error
Class FastConvergenceEstimator

java.lang.Object
  de.torstennahm.integrate.error.FastConvergenceEstimator

All Implemented Interfaces:

ErrorEstimator

public class FastConvergenceEstimator
extends java.lang.Object
implements ErrorEstimator

Provides an estimate based on the assumption that the integral values shows logarithmic convergence. By logarithmic convergence, we mean that log(error(n))=a-b*log(n), where n is the number of function evaluations, and a and b are parameters.

Author:

Torsten Nahm

Constructor Summary

FastConvergenceEstimator()
Create the convergence estimator with a default sample factor of 1.1.

FastConvergenceEstimator(double sampleFactor)
Create the convergence estimator with the specified sample factor.

Method Summary

double	getEstimate() Returns the estimate of the integration error based on the log.
double	getSlope()
void	log(long pointsEvaluated, double currentValue) Logs the current integration value.

Methods inherited from class java.lang.Object

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

Constructor Detail

FastConvergenceEstimator

public FastConvergenceEstimator()

Create the convergence estimator with a default sample factor of 1.1.

See Also:

FastConvergenceEstimator(double)

FastConvergenceEstimator

public FastConvergenceEstimator(double sampleFactor)

Create the convergence estimator with the specified sample factor. The sample factor determines how dense the interpolation points are to be placed. A factor of f means that when an interpolation point is set a n function evaluations, the next should be at f*n evaluations, and so on.

Parameters:

sampleFactor - factor between interpolation points

Method Detail

log

public void log(long pointsEvaluated,
                double currentValue)

Description copied from interface: ErrorEstimator

Logs the current integration value.

Specified by:

log in interface ErrorEstimator

Parameters:

pointsEvaluated - number of function evaluations for this integral value

currentValue - current integral value

getEstimate

public double getEstimate()

Description copied from interface: ErrorEstimator

Returns the estimate of the integration error based on the log.

Specified by:

getEstimate in interface ErrorEstimator

Returns:

estimate of the integration error

getSlope

public double getSlope()