Class Markowitz<T>

A search strategy based on methods outlined by Markowitz.

Inheritance

object

ParameterSet

Markowitz<T>

Implements

IParameterSet

ICloneable<Markowitz<T>>

Inherited Members

ParameterSet.SetParameter(string, P)

ParameterSet.TrySetParameter(string, P)

ParameterSet.GetProperty(string)

ParameterSet.TryGetProperty(string, out P)

ParameterSet.CreateParameterSetter(string)

ParameterSet.CreatePropertyGetter(string)

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Namespace: SpiceSharp.Algebra.Solve

Assembly: SpiceSharp.dll

Syntax

public class Markowitz<T> : ParameterSet, IParameterSet, ICloneable<Markowitz<T>>

Type Parameters

Name	Description
T	The base value type.

Constructors

Markowitz(Func<T, double>)

Initializes a new instance of the Markowitz<T> class.

Declaration

public Markowitz(Func<T, double> magnitude)

Parameters

Type	Name	Description
Func<T, double>	magnitude	The function for turning elements into a scalar.

Exceptions

Type	Condition
ArgumentNullException	Thrown if `magnitude` is `null`.

Properties

AbsolutePivotThreshold

Gets or sets the absolute threshold for choosing a pivot.

Declaration

[ParameterName("pivtol")]
[ParameterInfo("The absolute threshold for validating pivots")]
[GreaterThanOrEquals(0)]
[Finite]
public double AbsolutePivotThreshold { get; set; }

Property Value

Type	Description
double	The absolute pivot threshold.

Exceptions

Type	Condition
ArgumentOutOfRangeException	Thrown if the value is negative.

Magnitude

Gets the magnitude.

Declaration

public Func<T, double> Magnitude { get; }

Property Value

Type	Description
Func<T, double>	The magnitude.

RelativePivotThreshold

Gets or sets the relative threshold for choosing a pivot.

Declaration

[ParameterName("pivrel")]
[ParameterInfo("The relative threshold for validating pivots")]
[GreaterThan(0)]
[Finite]
public double RelativePivotThreshold { get; set; }

Property Value

Type	Description
double	The relative pivot threshold.

Exceptions

Type	Condition
ArgumentOutOfRangeException	Thrown if the value is not greater than 0.

Singletons

Gets the number of singletons.

Declaration

public int Singletons { get; }

Property Value

Type	Description
int

Strategies

Gets the strategies used for finding a pivot.

Declaration

public Collection<MarkowitzSearchStrategy<T>> Strategies { get; }

Property Value

Type	Description
Collection<MarkowitzSearchStrategy<T>>	The strategies.

Methods

Clear()

Clears the pivot strategy.

Declaration

public void Clear()

Clone()

Clones the instance.

Declaration

public Markowitz<T> Clone()

Returns

Type	Description
Markowitz<T>	The cloned instance.

ColumnCount(int)

Gets the Markowitz column counts.

Declaration

public int ColumnCount(int column)

Parameters

Type	Name	Description
int	column

Returns

Type	Description
int

Exceptions

Type	Condition
IndexOutOfRangeException	Thrown if `column` is negative or greater than the number of Markowitz columns.

CreateFillin(ISparseMatrix<T>, ISparseMatrixElement<T>)

Notifies the strategy that a fill-in has been created

Declaration

public void CreateFillin(ISparseMatrix<T> matrix, ISparseMatrixElement<T> fillin)

Parameters

Type	Name	Description
ISparseMatrix<T>	matrix	The matrix.
ISparseMatrixElement<T>	fillin	The fill-in.

Exceptions

Type	Condition
ArgumentNullException	Thrown if `matrix` or `fillin` is `null`.

FindPivot(ISparseMatrix<T>, int, int)

Find a pivot in the matrix.

Declaration

public Pivot<ISparseMatrixElement<T>> FindPivot(ISparseMatrix<T> matrix, int eliminationStep, int max)

Parameters

Type	Name	Description
ISparseMatrix<T>	matrix	The matrix.
int	eliminationStep	The current elimination step.
int	max	The maximum row/column index of any pivot.

Returns

Type	Description
Pivot<ISparseMatrixElement<T>>	The pivot information.

Remarks

The pivot should be searched for in the submatrix towards the right and down of the current diagonal at row/column eliminationStep. This pivot element will be moved to the diagonal for this elimination step.

Exceptions

Type	Condition
ArgumentNullException	Thrown if `matrix` is `null`.
ArgumentOutOfRangeException	Thrown if `eliminationStep` or `max` not 1 or higher, or `eliminationStep` is higher than `max`.

Initialize(IMatrix<T>)

Initializes the pivot searching algorithm.

Declaration

public void Initialize(IMatrix<T> matrix)

Parameters

Type	Name	Description
IMatrix<T>	matrix	The matrix to use for initialization.

Exceptions

Type	Condition
ArgumentNullException	Thrown if `matrix` is `null`.

IsValidPivot(ISparseMatrixElement<T>, int)

This method will check whether or not a pivot element is valid or not. It checks for the submatrix right/below of the pivot.

Declaration

public bool IsValidPivot(ISparseMatrixElement<T> pivot, int max)

Parameters

Type	Name	Description
ISparseMatrixElement<T>	pivot	The pivot candidate.
int	max	The maximum index that a pivot can have.

Returns

Type	Description
bool	True if the pivot can be used.

Exceptions

Type	Condition
ArgumentNullException	Thrown if `pivot` is `null`.

MovePivot(ISparseMatrix<T>, ISparseVector<T>, ISparseMatrixElement<T>, int)

Move the pivot to the diagonal for this elimination step.

Declaration

public void MovePivot(ISparseMatrix<T> matrix, ISparseVector<T> rhs, ISparseMatrixElement<T> pivot, int eliminationStep)

Parameters

Type	Name	Description
ISparseMatrix<T>	matrix	The matrix.
ISparseVector<T>	rhs	The right-hand side vector.
ISparseMatrixElement<T>	pivot	The pivot element.
int	eliminationStep	The elimination step.

Remarks

This is done by swapping the rows and columns of the diagonal and that of the pivot.

Exceptions

Type	Condition
ArgumentNullException	Thrown if `matrix`, `rhs` or `pivot` is `null`.

Product(int)

Gets the Markowitz products.

Declaration

public int Product(int index)

Parameters

Type	Name	Description
int	index

Returns

Type	Description
int

Exceptions

Type	Condition
IndexOutOfRangeException	Thrown if `index` is negative or greater than the number of Markowitz size.

RowCount(int)

Gets the Markowitz row counts.

Declaration

public int RowCount(int row)

Parameters

Type	Name	Description
int	row

Returns

Type	Description
int

Exceptions

Type	Condition
IndexOutOfRangeException	Thrown if `row` is negative or greater than the number of Markowitz rows.

Setup(ISparseMatrix<T>, ISparseVector<T>, int, int)

Setup the pivot strategy.

Declaration

public void Setup(ISparseMatrix<T> matrix, ISparseVector<T> rhs, int eliminationStep, int max)

Parameters

Type	Name	Description
ISparseMatrix<T>	matrix	The matrix.
ISparseVector<T>	rhs	The right-hand side vector.
int	eliminationStep	The current elimination step.
int	max	The maximum row/column index.

Exceptions

Type	Condition
ArgumentNullException	Thrown if `matrix` or `rhs` is `null`.

Update(ISparseMatrix<T>, ISparseMatrixElement<T>, int)

Update the strategy after the pivot was moved.

Declaration

public void Update(ISparseMatrix<T> matrix, ISparseMatrixElement<T> pivot, int limit)

Parameters

Type	Name	Description
ISparseMatrix<T>	matrix	The matrix.
ISparseMatrixElement<T>	pivot	The pivot element.
int	limit	The maximum row/column for pivots.

Exceptions

Type	Condition
ArgumentNullException	Thrown if `matrix` or `pivot` is `null`.

Implements

IParameterSet

ICloneable<T>

Extension Methods

Utility.ThrowIfNull<T>(T, string)

Documentation.ParameterValues<T>(IParameterSet, bool)

Documentation.Parameters(IParameterSet)