xal.tools.math.differential

Class DifferentiableOperationMinimizer

java.lang.Object

xal.tools.math.differential.DifferentiableOperationMinimizer

All Implemented Interfaces:

java.lang.Runnable

public class DifferentiableOperationMinimizer
extends java.lang.Object
implements java.lang.Runnable

Perform minimization on a differentiable operation

Constructor Summary

Constructors

Constructor and Description
DifferentiableOperationMinimizer(DifferentiableOperation operation, java.util.List<BoundedDifferentiableVariable> variables) Constructor with max evaluations of one
DifferentiableOperationMinimizer(DifferentiableOperation operation, java.util.List<BoundedDifferentiableVariable> variables, int maxEvaluations) Primary Constructor

Method Summary

Modifier and Type	Method and Description
xal.tools.math.differential.Trial	findNextRandomTrialWithPowerDistribution(xal.tools.math.differential.Trial initialTrial, int evaluations) find the next random trial with the default power distribution
xal.tools.math.differential.Trial	findNextRandomTrialWithPowerDistribution(xal.tools.math.differential.Trial initialTrial, int evaluations, double power) find the next random trial with the specified power distribution
xal.tools.math.differential.Trial	findNextTrial(xal.tools.math.differential.Trial initialTrial, int evaluations) determine the next trial to evaluate
xal.tools.math.differential.Trial	findNextTrialAlongGradient(xal.tools.math.differential.Trial initialTrial, int evaluations, double step, double[] gradient) find the next trial taking the specified step along the gradient
double	getBestPenalty() Get the penalty for the best solution
double	getBestVariableValue(BoundedDifferentiableVariable variable) Get the best solution value for the specified variable
DifferentiableVariableValues	getBestVariableValueMap() Get the best solution variable value map
static DifferentiableOperationMinimizer	getInstance(DifferentiableOperation operation, BoundedDifferentiableVariable... variables) Construct a minimizer from the operation and the variables to explore
int	getMaxEvaluations() Get the maximum number of iterations to evaluate the operation
double	getMaxSearchStep(xal.tools.math.differential.TrialPoint initialPoint, double[] gradient) get the maximum search step along the gradient
double	getPenalty(xal.tools.math.differential.TrialPoint point) evaluate the operation to minimize at the specified trial point
void	reset() reset this minimizer
void	run() run the optimization
void	run(int maxEvaluations) set the maximum number of evaluations and run the optimization
void	setMaxEvaluations(int evaluations) set the maximum number of evaluations of the penalty operation

Methods inherited from class java.lang.Object

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

Constructor Detail

DifferentiableOperationMinimizer

public DifferentiableOperationMinimizer(DifferentiableOperation operation,
                                        java.util.List<BoundedDifferentiableVariable> variables,
                                        int maxEvaluations)

Primary Constructor

DifferentiableOperationMinimizer

public DifferentiableOperationMinimizer(DifferentiableOperation operation,
                                        java.util.List<BoundedDifferentiableVariable> variables)

Constructor with max evaluations of one

Method Detail

getInstance

public static DifferentiableOperationMinimizer getInstance(DifferentiableOperation operation,
                                                           BoundedDifferentiableVariable... variables)

Construct a minimizer from the operation and the variables to explore

reset

public void reset()

reset this minimizer

setMaxEvaluations

public void setMaxEvaluations(int evaluations)

set the maximum number of evaluations of the penalty operation

getMaxEvaluations

public int getMaxEvaluations()

Get the maximum number of iterations to evaluate the operation

run

public void run()

run the optimization

Specified by:

run in interface java.lang.Runnable

run

public void run(int maxEvaluations)

set the maximum number of evaluations and run the optimization

getBestPenalty

public double getBestPenalty()

Get the penalty for the best solution

getBestVariableValue

public double getBestVariableValue(BoundedDifferentiableVariable variable)

Get the best solution value for the specified variable

getBestVariableValueMap

public DifferentiableVariableValues getBestVariableValueMap()

Get the best solution variable value map

getPenalty

public double getPenalty(xal.tools.math.differential.TrialPoint point)

evaluate the operation to minimize at the specified trial point

getMaxSearchStep

public double getMaxSearchStep(xal.tools.math.differential.TrialPoint initialPoint,
                               double[] gradient)

get the maximum search step along the gradient

findNextTrial

public xal.tools.math.differential.Trial findNextTrial(xal.tools.math.differential.Trial initialTrial,
                                                       int evaluations)

determine the next trial to evaluate

findNextTrialAlongGradient

public xal.tools.math.differential.Trial findNextTrialAlongGradient(xal.tools.math.differential.Trial initialTrial,
                                                                    int evaluations,
                                                                    double step,
                                                                    double[] gradient)

find the next trial taking the specified step along the gradient

findNextRandomTrialWithPowerDistribution

public xal.tools.math.differential.Trial findNextRandomTrialWithPowerDistribution(xal.tools.math.differential.Trial initialTrial,
                                                                                  int evaluations)

find the next random trial with the default power distribution

findNextRandomTrialWithPowerDistribution

public xal.tools.math.differential.Trial findNextRandomTrialWithPowerDistribution(xal.tools.math.differential.Trial initialTrial,
                                                                                  int evaluations,
                                                                                  double power)

find the next random trial with the specified power distribution