Package pulse.search.linear

Class LinearOptimiser

java.lang.Object

pulse.util.UpwardsNavigable

pulse.util.Group

pulse.util.Accessible

pulse.util.PropertyHolder

pulse.search.linear.LinearOptimiser

All Implemented Interfaces:

Serializable, Descriptive, Reflexive

Direct Known Subclasses:

GoldenSectionOptimiser, WolfeOptimiser

public abstract class LinearOptimiser
extends PropertyHolder
implements Reflexive

The most basic LinearSolver class, which defines the notion of the linear resolution, defines the method signature for estimating the step of the linear search (i.e., the position of the minimum), and provides a simple algorithm to initialise the calculation domain.

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field	Description
protected static double	searchResolution

Constructor Summary

Constructors

Modifier	Constructor	Description
protected	LinearOptimiser()

Method Summary

Modifier and Type	Method	Description
static Segment	domain(ParameterVector x, Vector p)	Sets the domain for this linear search on p.
static NumericProperty	getLinearResolution()	The linear resolution determines the minimum distance between any two points belonging to the domain of this search while they still are considered separate.
abstract double	linearStep(GeneralTask task)	Finds the minimum of the target function on the domain Segment.
Set<NumericPropertyKeyword>	listedKeywords()	The LINEAR_RESOLUTION is the single listed parameter for this class.
void	set(NumericPropertyKeyword type, NumericProperty property)	An abstract method, which must be overriden to gain access over setting the values of all relevant (selected by the programmer) NumericProperties in subclasses of Accessible.
static void	setLinearResolution(NumericProperty searchError)
String	toString()

Methods inherited from class pulse.util.PropertyHolder

addListener, areDetailsHidden, data, describe, firePropertyChanged, getDescriptor, getListeners, getPrefix, ignoreSiblings, initListeners, isListedNumericType, isListedParameter, listedTypes, numericData, parameterListChanged, removeListeners, setPrefix, updateProperties, updateProperty

Methods inherited from class pulse.util.Accessible

accessibleChildren, genericProperties, genericProperty, numericProperties, numericProperty, property, update, update

Methods inherited from class pulse.util.Group

access, children, contents, getSimpleName, subgroups

Methods inherited from class pulse.util.UpwardsNavigable

addHierarchyListener, getHierarchyListeners, getParent, identify, removeHierarchyListener, removeHierarchyListeners, setParent, specificAncestor, tellParent

Methods inherited from class java.lang.Object

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

Field Detail

searchResolution

protected static double searchResolution

Constructor Detail

LinearOptimiser

protected LinearOptimiser()

Method Detail

linearStep

public abstract double linearStep(GeneralTask task)
                           throws SolverException

Finds the minimum of the target function on the domain Segment.

Parameters:

task - the target function is the sum of squared residuals (SSR) for this task

Returns:

a double, representing the step magnitude that needs to be multiplied by the direction of the search determined previously using the PathSolver to arrive at the next set of parameters corresponding to a lower SSR value of this task

Throws:

SolverException

domain

public static Segment domain(ParameterVector x,
                             Vector p)

Sets the domain for this linear search on p.

The domain is defined as a Segment [0; max], where max determines the maximum magnitude of the linearStep. This value is calculated initially as max = 0.5*xi/pi, where i is the index of the DIFFUSIVITY NumericProperty. Later it is corrected to ensure that the change in the HEAT_LOSS NumericProperty is less than unity.

Parameters:

x - the current set of parameter

p - the result of the direction search with the PathSolver

Returns:

a Segment defining the domain of this search

getLinearResolution

public static NumericProperty getLinearResolution()

The linear resolution determines the minimum distance between any two points belonging to the domain of this search while they still are considered separate. In case of a partitioning method, e.g. the golden-section search, this determines the partitioning limit. Note different PathSolvers can have different sensitivities to the linear search and may require different linear resolutions to work effectively.

Returns:

a NumericProperty with the current value of the linear resolution

See Also:

domain(IndexedVector,IndexedVector,Vector)

setLinearResolution

public static void setLinearResolution(NumericProperty searchError)

toString

public String toString()

Overrides:

toString in class Object

listedKeywords

public Set<NumericPropertyKeyword> listedKeywords()

The LINEAR_RESOLUTION is the single listed parameter for this class.

Overrides:

listedKeywords in class PropertyHolder

Returns:

a set of NumericPropertyKeyword instances, which have been explicitly marked as a listed parameter for this PropertyHolder.

See Also:

NumericPropertyKeyword

set

public void set(NumericPropertyKeyword type,
                NumericProperty property)

Description copied from class: Accessible

An abstract method, which must be overriden to gain access over setting the values of all relevant (selected by the programmer) NumericProperties in subclasses of Accessible. Typically this involves a switch statement that goes through the different options for the type and invokes different set(...) methods to update the matching NumericProperty with property.

Specified by:

set in class Accessible

Parameters:

type - the type, which must be equal by definition to property.getType().

property - the property, which contains new information.