14.2.42 Class Variable¶

mosek::fusion::Variable¶

An abstract variable object. This is the base class for all variable types in Fusion, and it contains several methods for manipulating variable objects.

The Variable object can be an interface to the normal model variables, e.g. LinearVariable and ConicVariable, to slices of other variables or to concatenations of other variables.

Primal and dual solution values can be accessed through the Variable object.

More static methods to manipulate variables are available in the Var class.

Implements

Expression

Members

Expression.eval – Evaluate the expression and push the result onto the work stack.

Expression.getDim – Return the d’th dimension in the expression.

Variable.antidiag – Return the anti-diagonal of a variable matrix.

Variable.asExpr – Create an expression corresponding to the variable object.

Variable.diag – Return the diagonal of a variable matrix.

Variable.dual – Get the dual solution value of the variable.

Variable.fromTril – Convert from a linear representation of the lower triangular part of a square variable into a square variable.

Variable.getModel – Return the model to which the variable belongs.

Variable.getND – Get the number of dimensions in the variable shape.

Variable.getShape – Return the shape of the variable.

Variable.getSize – Get the total number of elements in the variable.

Variable.index – Return a single entry in the variable.

Variable.level – Return the primal value of the variable as an array.

Variable.makeContinuous – Drop integrality constraints on the variable, if any.

Variable.makeInteger – Apply integrality constraints on the variable. Has no effect on elements of semidefinite matrix variables.

Variable.pick – Create a one-dimensional variable by picking a list of indexes from this variable.

Variable.remove – Remove the variable from the model.

Variable.reshape – Reshape the variable. The new shape must have the same total size as the current.

Variable.setLevel – Input solution values for this variable

Variable.slice – Create a slice variable by picking a range of indexes for each variable dimension.

Variable.toString – Create a string representation of the variable.

Variable.transpose – Transpose the variable.

Variable.tril – Convert a square variable into a linear representation of the lower triangular part of the variable.

Implemented by

BaseVariable

Variable.antidiag¶

Variable::t antidiag(int index)
Variable::t antidiag()

Return the anti-diagonal of a variable matrix in a one-dimensional variable object. The main anti-diagonal is defined as starting with the element in the first row and last column.

Parameters: index (int) – Index of the anti-diagonal. Index 0 means the main anti-diagonal, negative indices are below it and positive indices are above it.
Return: (Variable)

Variable.asExpr¶

Expression::t asExpr()

Create an Expression object corresponding to \(I\cdot V\), where \(I\) is the identity matrix and \(V\) is this variable.

Return: (Expression)

Variable.diag¶

Variable::t diag(int index)
Variable::t diag()

Return the diagonal of a square variable matrix in a one-dimensional variable object. The main diagonal is defined as starting with the element in the first row and first column.

Parameters: index (int) – Index of the diagonal. Index 0 means the main diagonal, negative indices are below it and positive indices are above it.
Return: (Variable)

Variable.dual¶

shared_ptr<ndarray<double,1>> dual()

Get the dual solution value of the variable as an array. When the selected slice is multi-dimensional, this corresponds to the flattened slice of solution values.

Return: (double[])

Variable.fromTril¶

Variable::t fromTril(int dim)

Convert from a linear representation of the lower triangular part of a square variable into a square variable.

Parameters: dim (int) – Dimension of the square variable.
Return: (Variable)

Variable.getModel¶

Model::t getModel()

Get the Model object that the variable belongs to.

Return: (Model)

Variable.getND¶

int getND()

Get the number of dimensions in the variable shape.

Return: (int)

Variable.getShape¶

shared_ptr<ndarray<int,1>> getShape()

Get the variable shape.

Return: (int[])

Variable.getSize¶

long long getSize()

Get the total number of elements in the variable.

Return: (long long)

Variable.index¶

Variable::t index(int i1)
Variable::t index(int i1, int i2)
Variable::t index(int i1, int i2, int i3)
Variable::t index(shared_ptr<ndarray<int,1>> idx)

Return a variable of size one corresponding to a single element in the variable object.

Parameters

i1 (int) – Index in the first dimension of the element requested.
i2 (int) – Index in the second dimension of the element requested.
i3 (int) – Index in the third dimension of the element requested.
idx (int[]) – List of indexes of the elements requested.

Return

(Variable)

Variable.level¶

shared_ptr<ndarray<double,1>> level()

Get the primal solution value of the variable as an array. When the selected slice is multi-dimensional, this corresponds to the flattened slice of solution values.

Return: (double[])

Variable.makeContinuous¶

void makeContinuous()

Drop integrality constraints on the variable, if any.

Variable.makeInteger¶

void makeInteger()

Apply integrality constraints on the variable. Has no effect on elements of semidefinite matrix variables.

Variable.pick¶

Variable::t pick(shared_ptr<ndarray<int,1>> idxs)
Variable::t pick(shared_ptr<ndarray<int,2>> midxs)
Variable::t pick(shared_ptr<ndarray<int,1>> i1, shared_ptr<ndarray<int,1>> i2)
Variable::t pick(shared_ptr<ndarray<int,1>> i1, shared_ptr<ndarray<int,1>> i2, shared_ptr<ndarray<int,1>> i3)

Create a one-dimensional variable by picking a list of indexes from this variable.

Parameters

idxs (int[]) – Indexes of the elements requested.
midxs (int[][]) – A sequence of multi-dimensional indexes of the elements requested.
i1 (int[]) – Index along the first dimension.
i2 (int[]) – Index along the second dimension.
i3 (int[]) – Index along the third dimension.

Return

(Variable)

Variable.remove¶

void remove()

Remove the variable from the model and remove it from any constraints where it appears. Using the variable object after this method has been called results in undefined behavior.

Variable.reshape¶

Variable::t reshape(shared_ptr<ndarray<int,1>> shape)
Variable::t reshape(int dim0)
Variable::t reshape(int dim0, int dim1)
Variable::t reshape(int dim0, int dim1, int dim2)

Reshape the variable. The new shape must have the same total size as the current.

Parameters

shape (int[]) – The new shape.
dim0 (int) – First dimension of new shape
dim1 (int) – Second dimension of new shape
dim2 (int) – Third dimension of new shape

Return

(Variable)

Variable.setLevel¶

void setLevel(shared_ptr<ndarray<double,1>> v)

Set values for an initial solution for this variable. Note that these values are buffered until the solver is called; they are not available through the level() methods.

Parameters: v (double[]) – An array of values to be assigned to the variable.

Variable.slice¶

Variable::t slice(int first, int last)
Variable::t slice(shared_ptr<ndarray<int,1>> firsta, shared_ptr<ndarray<int,1>> lasta)

Create a slice variable by picking a range of indexes for each variable dimension.

Parameters

first (int) – The index from which the slice begins.
last (int) – The index after the last element of the slice.
firsta (int[]) – The indices from which the slice of a multidimensional variable begins.
lasta (int[]) – The indices after the last element of slice of a multidimensional variable.

Return

(Variable)

Variable.toString¶

string toString()

Create a string representation of the variable.

Return: (string)

Variable.transpose¶

Variable::t transpose()

Return the transpose of the current variable. The variable must have at most two dimensions.

Return: (Variable)

Variable.tril¶

Variable::t tril()

Convert a square variable into a linear representation of the lower triangular part of the variable.

Return: (Variable)

14.2.42 Class Variable¶

Table of Contents

Download PDF

Modeling Cookbook

Cheatsheet