Package libai.common.matrix

Class Matrix

java.lang.Object

libai.common.matrix.Matrix

All Implemented Interfaces:

java.io.Serializable

Direct Known Subclasses:

Column, Row

public class Matrix
extends java.lang.Object
implements java.io.Serializable

Matrix implementation. This class can handle the basic operations of matrix space, add, subtract, product scalar product, transpose, and other useful operations for this API. For performance reasons, precondition validations are done using asserts instead of the more clear and safe exceptions.

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field	Description
protected int	columns	Number of rows and columns of the matrix.
protected double[]	matrix	Matrix's data, stored for row in a sequential array.
protected int	rows	Number of rows and columns of the matrix.

Constructor Summary

Constructors

Constructor	Description
Matrix(int r, int c)	Constructor alias for Matrix(r,c,false).
Matrix(int r, int c, boolean identity)	Constructor.
Matrix(int r, int c, double[] data)	Constructor.
Matrix(Matrix copy)	Copy constructor.

Method Summary

Modifier and Type	Method	Description
void	add(Matrix a, Matrix b)	Adds two matrices.
void	apply(Function f, Matrix a)	Apply one function over the elements of the matrix and left the result on a.
void	applyInIdentity(Function f, Matrix a)	Apply one function over the elements of the matrix and left on the main diagonal of a identity matrix 'a'.
void	copy(Matrix a)	Copy this matrix to another matrix a.
double	dotProduct(Matrix a)	Calculate the dot product between this matrix and another.
boolean	equals(java.lang.Object b1)	Check if two matrix are equals position to position with a precision of 1e-7.
void	fill()	Fill the matrix with random values between (-1, 1).
void	fill(boolean signed)	Fill the matrix with random values between [0, 1) if signed is is false, and (-1, 1) if true.
void	fill(boolean signed, java.util.Random r)	Fill the matrix with random values between [0, 1) if signed is is false, and (-1, 1) if true.
double[]	getCol(int index)	Return an array with the values of the specified column.
int	getColumns()
static java.util.Random	getDefaultRandom()
double[]	getRow(int index)	Return an array with the values of the specified row.
int	getRows()
int	hashCode()
void	increment(int i, int j, double v)	Increments the value of one position by v.
void	multiply(double a, Matrix b)	Multiply this matrix by an scalar.
void	multiply(Matrix a, Matrix b)	Multiply two matrix.
void	multiplyAndAdd(double a, Matrix b, Matrix result)	Return in result the value of (this*a + b) for each This(i,j)
double	position(int i, int j)	Return the value on the position (i,j).
void	position(int i, int j, double v)	Set the value v on the position (i,j).
static Matrix	random(int r, int c)	Create a new Matrix filled with low random numbers.
static Matrix	random(int r, int c, boolean signed)	Create a new Matrix filled with low random numbers.
static Matrix	random(int r, int c, boolean signed, java.util.Random rand)	Create a new Matrix filled with low random numbers.
void	scale(int i, int j, double v)	Scales the value of one position by v.
void	setRow(int index, double[] data)	Replace one row of values of this matrix.
void	setValue(double v)	Set a value on each position of the matrix.
Column	solve(Column rhs)	Return the reduced echelon form of this matrix using the argument as a right hand side colum.
void	subtract(Matrix a, Matrix b)	Subtract two matrices.
void	subtractAndCopy(Matrix b, Matrix resultSubtract, Matrix resultCopy)	Subtract the value of this with the value of b, and let the result on resultSubstract Also, copy the original value of this into resultCopy.
void	swap(int i1, int i2)	Swaps two rows.
java.lang.String	toString()	Return the string representation of this matrix.
java.lang.String	toString(int precision)
Matrix	transpose()	Return the transpose of this matrix.
void	transpose(Matrix a)	Transpose this matrix and left the result on a.

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Field Detail

matrix

protected final double[] matrix

Matrix's data, stored for row in a sequential array.

rows

protected final int rows

Number of rows and columns of the matrix.

columns

protected final int columns

Number of rows and columns of the matrix.

Constructor Detail

Matrix

public Matrix(int r,
              int c,
              boolean identity)

Constructor. Allocated the matrix and could initialize with the identity. By default, the matrix is created filled with zeroes.

Parameters:

r - number of rows

c - number of columns

identity - if you need initialized with an identity.

Throws:

java.lang.IllegalArgumentException - if either r or c are less or equal than zero.

Matrix

public Matrix(int r,
              int c)

Constructor alias for Matrix(r,c,false).

Parameters:

r - number of rows

c - number of columns

Throws:

java.lang.IllegalArgumentException - if either r or c are less or equal than zero.

Matrix

public Matrix(int r,
              int c,
              double[] data)

Constructor.

Creates a matrix and initialize with the data on data.

The values are read as row based: data -> row1, row2, ..., rown

Parameters:

r - number of rows

c - number of columns

data - values to initialize the matrix (length must be r * c).

Throws:

java.lang.IllegalArgumentException - if either r or c are less or equal than zero or if data is null or has the wrong dimension.

Matrix

public Matrix(Matrix copy)

Copy constructor. Creates a copy of a matrix using another matrix. Constructor alias for Matrix(copy.rows, copy.columns, copy.matrix).

Parameters:

copy - Source of the copy

Method Detail

getDefaultRandom

public static java.util.Random getDefaultRandom()

random

public static Matrix random(int r,
                            int c,
                            boolean signed)

Create a new Matrix filled with low random numbers.

Parameters:

r - number of rows

c - number of columns

signed - true if the matrix should be filled with positive and negative numbers false if the numbers should be greater or equal than zero.

Returns:

a new matrix filled with low random numbers.

Throws:

java.lang.IllegalArgumentException - if either r or c are less or equal than zero.

See Also:

fill(boolean, java.util.Random)

random

public static Matrix random(int r,
                            int c,
                            boolean signed,
                            java.util.Random rand)

Create a new Matrix filled with low random numbers.

Parameters:

r - number of rows

c - number of columns

signed - true if the matrix should be filled with positive and negative numbers false if the numbers should be greater or equal than zero.

rand - The Random object used to fill the matrix, if null it will fallback to ThreadLocalRandom.current()

Returns:

a new matrix filled with low random numbers.

Throws:

java.lang.IllegalArgumentException - if either r or c are less or equal than zero.

See Also:

fill(boolean, java.util.Random)

random

public static Matrix random(int r,
                            int c)

Create a new Matrix filled with low random numbers.

Parameters:

r - number of rows

c - number of columns

Returns:

a new matrix filled with low random numbers.

Throws:

java.lang.IllegalArgumentException - if either r or c are less or equal than zero.

See Also:

random(int, int, boolean), random(int, int, boolean, java.util.Random)

add

public void add(Matrix a,
                Matrix b)

Adds two matrices. b = this + a

The matrix b must be created and has the same dimension of this and a.

NOTE: Assertions of the dimensions are made with assert statement. You must enable this on runtime to be effective.

Parameters:

a - The matrix to add

b - The matrix to put the result

subtract

public void subtract(Matrix a,
                     Matrix b)

Subtract two matrices. b = this - a.

The matrix b must be created and has the same dimension of this and a.

NOTE: Assertions of the dimensions are made with assert statement. You must enable this on runtime to be effective.

Parameters:

a - The matrix to subtract

b - The matrix to put the result

multiply

public void multiply(double a,
                     Matrix b)

Multiply this matrix by an scalar. b = this * a.

The matrix b must be created and has the same dimension of this.

NOTE: Assertions of the dimensions are made with assert statement. You must enable this on runtime to be effective.

Parameters:

a - The scalar to multiply

b - The matrix to put the result

multiply

public void multiply(Matrix a,
                     Matrix b)

Multiply two matrix. b = this * a

The matrix a must be created and has the right dimensions, that is a.rows = this.columns.

The matrix b must be created and has the right dimensions, that is b.rows = this.rows and b.columns = a.columns.

NOTE: Assertions of the dimensions are made with assert statement. You must enable this on runtime to be effective.

Parameters:

a - The matrix to multiply

b - The matrix to put the result

apply

public void apply(Function f,
                  Matrix a)

Apply one function over the elements of the matrix and left the result on a. For each element on this a(i,j) = F(this(i,j)).

The matrix a must have the same dimension as this.

NOTE: Assertions of the dimensions are made with assert statement. You must enable this on runtime to be effective.

Parameters:

f - function to apply.

a - The matrix to put the result.

applyInIdentity

public void applyInIdentity(Function f,
                            Matrix a)

Apply one function over the elements of the matrix and left on the main diagonal of a identity matrix 'a'. For each element on this a(i,i) = F(this(i,j)). The matrix this must be a column matrix. The matrix a must be created and has the same dimension of this and a. NOTE: Assertions of the dimensions are made with assert statement. You must enable this on runtime to be effective.

Parameters:

f - function to apply.

a - The matrix to put the result.

fill

public void fill()

Fill the matrix with random values between (-1, 1). Alias for fill(true).

See Also:

fill(boolean), fill(boolean, java.util.Random)

fill

public void fill(boolean signed)

Fill the matrix with random values between [0, 1) if signed is is false, and (-1, 1) if true.

Parameters:

signed - false if all the numbers should be positive, false otherwise

See Also:

fill(boolean, java.util.Random)

fill

public void fill(boolean signed,
                 java.util.Random r)

Fill the matrix with random values between [0, 1) if signed is is false, and (-1, 1) if true.

This method is based only in Random.nextDouble(), so in case other intervals are needed the only thing that's needed is a custom implementation of nextDouble(), for instance:

     Random myRand = new Random(){
         public double nextDouble() {
             return super.nextDouble() / 1000.;
         }
     }
 

Parameters:

signed - false if all the numbers should be positive, false otherwise

r - The Random object used to fill the matrix, if null it will fallback to ThreadLocalRandom.current()

copy

public void copy(Matrix a)

Copy this matrix to another matrix a.

The matrix a must have the same dimension as this.

NOTE: Assertions of the dimensions are made with assert statement. You must enable this on runtime to be effective.

Parameters:

a - The matrix to put the result.

dotProduct

public double dotProduct(Matrix a)

Calculate the dot product between this matrix and another. This method can safely calculate the dot product between row-column, row-row, column-column matrices.

Parameters:

a - The matrix to multiply

Returns:

the scalar of the dot product.

Throws:

java.lang.IllegalArgumentException - if either this or a are neither row nor column matrices.

transpose

public void transpose(Matrix a)

Transpose this matrix and left the result on a.

The matrix a must be created and has the right dimensions.

NOTE: Assertions of the dimensions are made with assert statement. You must enable this on runtime to be effective.

Parameters:

a - The matrix to put the result.

transpose

public Matrix transpose()

Return the transpose of this matrix.

Returns:

A new matrix with the transpose of this.

setRow

public void setRow(int index,
                   double[] data)

Replace one row of values of this matrix.

The number of values of data must match with the number of columns of this.

NOTE: Assertions of the dimensions and indexes are made with assert statement. You must enable this on runtime to be effective.

Parameters:

index - The index of the row to place the values.

data - The values to put in that row.

setValue

public void setValue(double v)

Set a value on each position of the matrix.

Parameters:

v - The value to put in the matrix.

getRow

public double[] getRow(int index)

Return an array with the values of the specified row.

NOTE: Assertions of the indexes are made with assert statement. You must enable this on runtime to be effective.

Parameters:

index - The index of the row to return.

Returns:

the array with the values.

getCol

public double[] getCol(int index)

Return an array with the values of the specified column.

NOTE: Assertions of the indexes are made with assert statement. You must enable this on runtime to be effective.

Parameters:

index - The index of the column to return.

Returns:

the array with the values.

position

public final double position(int i,
                             int j)

Return the value on the position (i,j).

NOTE: Assertions of the indexes are made with assert statement. You must enable this on runtime to be effective.

Parameters:

i - index of the row

j - index of the column

Returns:

the value of that position

position

public final void position(int i,
                           int j,
                           double v)

Set the value v on the position (i,j).

NOTE: Assertions of the indexes are made with assert statement. You must enable this on runtime to be effective.

Parameters:

i - index of the row

j - index of the column

v - the value to put into.

subtractAndCopy

public void subtractAndCopy(Matrix b,
                            Matrix resultSubtract,
                            Matrix resultCopy)

Subtract the value of this with the value of b, and let the result on resultSubstract Also, copy the original value of this into resultCopy.

Parameters:

b - Matrix to subtract

resultSubtract - Matrix to hold the result of the subtraction

resultCopy - Matrix to hold the copy of this.

multiplyAndAdd

public void multiplyAndAdd(double a,
                           Matrix b,
                           Matrix result)

Return in result the value of (this*a + b) for each This(i,j)

Parameters:

a - constant to multiply

b - matrix to add

result - Matrix to hold the result of the operation.

increment

public final void increment(int i,
                            int j,
                            double v)

Increments the value of one position by v.

NOTE: Assertions of the indexes are made with assert statement. You must enable this on runtime to be effective.

Parameters:

i - index of the row

j - index of the column

v - value to increment.

scale

public final void scale(int i,
                        int j,
                        double v)

Scales the value of one position by v.

NOTE: Assertions of the indexes are made with assert statement. You must enable this on runtime to be effective.

Parameters:

i - index of the row

j - index of the column

v - value to scale.

swap

public void swap(int i1,
                 int i2)

Swaps two rows.

NOTE: Assertions of the indexes are made with assert statement. You must enable this on runtime to be effective.

Parameters:

i1 - index of the first row

i2 - index of the second row

solve

public Column solve(Column rhs)

Return the reduced echelon form of this matrix using the argument as a right hand side colum.

Parameters:

rhs - Column to use as right hand side

Returns:

rref([this rhs])(:,end)

Since:

1.2.0-RIDDLER

equals

public boolean equals(java.lang.Object b1)

Check if two matrix are equals position to position with a precision of 1e-7. If the dimensions mismatch they aren't equals. If one position differs by more than 1e-7 then are different.

Overrides:

equals in class java.lang.Object

Parameters:

b1 - The matrix to compare

Returns:

true if are equals, false otherwise.

hashCode

public int hashCode()

Overrides:

hashCode in class java.lang.Object

toString

public java.lang.String toString()

Return the string representation of this matrix. Useful to write on a file or for debugging.

Overrides:

toString in class java.lang.Object

Returns:

An string with the values of the matrix.

toString

public java.lang.String toString(int precision)

getRows

public int getRows()

Returns:

The number of rows

getColumns

public int getColumns()

Returns:

The number of columns