lumix.nonlinear.terms.LXAbsoluteTerm

class lumix.nonlinear.terms.LXAbsoluteTerm(var, coefficient=1.0)[source]

Absolute value term: x.

Represents the absolute value of a variable, which can be linearized using auxiliary variables and linear constraints. The linearization introduces a new variable z and constraints: z >= x and z >= -x.

Parameters:
var

The variable to take the absolute value of.

coefficient

Coefficient to multiply the absolute value by (default: 1.0).

Example

Basic absolute value term:

from lumix.nonlinear import LXAbsoluteTerm
from lumix.core import LXVariable

# Minimize absolute deviation from target
actual = LXVariable[Product, float]("actual").from_data(products)
abs_deviation = LXAbsoluteTerm(var=actual, coefficient=1.0)

With coefficient:

# Weighted absolute penalty
penalty = LXAbsoluteTerm(var=deviation, coefficient=10.0)

Note

The linearization creates auxiliary variables and adds constraints automatically during the model building phase. The auxiliary variable z will appear in the objective function or constraints where the absolute value is used.

__init__(var, coefficient=1.0)
Parameters:
Return type:

None

Methods

__init__(var[, coefficient])

Attributes

coefficient

var

var: LXVariable
coefficient: float = 1.0
__init__(var, coefficient=1.0)
Parameters:
Return type:

None