VariableGenerator

class VariableGenerator

Provides functionality to generate a decision variable or an array of decision variables.

Methods

__init__

array

Generate an array of variables.

matrix

Generate a quadratic polynomial in coefficient-matrix form.

scalar

Generate a scalar variable.

Attributes

variables

List of all variables created.
__init__(self) → None
__repr__(self) → str

Return a developer-friendly string representation of the generator.

Return type:

str

__str__(self) → str

Return a human-readable string representation of the generator.

Return type:

str

array(
self,
type: Literal['Binary', 'Ising', 'Integer', 'Real'] | VariableType,
shape: tuple[int, ...] | int,
bounds: tuple[float | None, ...] = (None, None),
name: str = '',
) → PolyArray[Dim]
array(
self,
type: Literal['Binary', 'Ising', 'Integer', 'Real'] | VariableType,
*shape: int,
) → PolyArray[Dim]

Generate an array of variables.

Variables are generated in an n-dimensional array. Each variable has the same type and bounds (if any).

Overload 1.

Args:

  • type (typing.Literal[‘Binary’, ‘Ising’, ‘Integer’, ‘Real’] | amplify.VariableType): The type of variables.

  • shape (tuple[int, …] | int): The NumPy-like shape of the variable array.

    For example, (3, 4) generates a two-dimensional variable array with 3 rows and 4 columns.

  • bounds (tuple[float | None, …]): The lower and upper bounds of the variables.

    If either of the bounds is not specified, it will be unbounded in that direction. A Binary or an Ising variable carries no bound. Giving one that narrows its value set raises a ValueError. Defaults to (None, None).

  • name (str): The name of the variable array.

    Each variable is automatically assigned a name consisting of this string and a specific index. Defaults to ''.

Returns:

amplify.PolyArray[Dim]: Generated variable array.

An array of the given shape. Every element is a polynomial of one new variable.

Overload 2.

Args:

  • type (typing.Literal[‘Binary’, ‘Ising’, ‘Integer’, ‘Real’] | amplify.VariableType): The type of variables.

  • *shape (int): The NumPy-like shape of the variable array, given as one integer for each axis.

Returns:

amplify.PolyArray[Dim]: Generated variable array.

matrix(
self,
type: Literal['Binary', 'Ising', 'Integer', 'Real'] | VariableType,
shape: int | tuple[int, ...],
bounds: tuple[float | None, ...] = (None, None),
name: str = '',
) → Matrix
matrix(
self,
type: Literal['Binary', 'Ising', 'Integer', 'Real'] | VariableType,
*shape: int,
) → Matrix

Generate a quadratic polynomial in coefficient-matrix form.

A variable array of the specified shape is generated together with the coefficient matrix. The polynomial represented by the coefficient matrix is defined as \((x^T) Q x + (p^T) x + c\), where Q is the quadratic coefficient matrix, p is the linear coefficient array, c is the constant term, and x is the variable array.

Overload 1.

Args:

  • type (typing.Literal[‘Binary’, ‘Ising’, ‘Integer’, ‘Real’] | amplify.VariableType): The type of variables.

  • shape (int | tuple[int, …]): The shape of the array. It can be specified in two ways:

    • A single integer, e.g., 5, which will generate a one-dimensional variable array with 5 elements.

    • A NumPy-like shape, e.g., (3, 4), which will generate a two-dimensional variable array with 3 rows and 4 columns.

  • bounds (tuple[float | None, …]): The lower and upper bounds of the variables.

    If either bound is omitted, the variable is unbounded in that direction. A Binary or an Ising variable carries no bound. Giving one that narrows its value set raises a ValueError. Defaults to (None, None).

  • name (str): The name of the variable array.

    Each variable is automatically assigned a name consisting of this string and a specific index. Defaults to ''.

Returns:

amplify.Matrix: Coefficient matrix.

A matrix whose coefficients are all zero. Its variable_array holds the new variables that the matrix uses.

Overload 2.

Args:

  • type (typing.Literal[‘Binary’, ‘Ising’, ‘Integer’, ‘Real’] | amplify.VariableType): The type of variables.

  • *shape (int): The shape of the array.

    This overload is provided for backward compatibility.

Returns:

amplify.Matrix: Coefficient matrix.

scalar(
self,
type: Literal['Binary', 'Ising', 'Integer', 'Real'] | VariableType,
bounds: tuple[float | None, ...] = (None, None),
name: str = '',
) → Poly

Generate a scalar variable.

The variable will have the specified type and bounds (if any).

Parameters:

  • type (Literal['Binary', 'Ising', 'Integer', 'Real'] | amplify.VariableType) – The type of the variable.

  • bounds (tuple[float | None, ...]) – The lower and upper bounds of the variable. A Binary or an Ising variable carries no bound. Giving one that narrows its value set raises a ValueError. Defaults to (None, None).

  • name (str) – The name of the variable. Defaults to ''.

Returns:

Generated variable.

A polynomial of one new variable.

Return type:

Poly

property variables

List of all variables created.

Return type:

list[Variable]