csnlp.multistart.RandomStartPoints#

class csnlp.multistart.RandomStartPoints(points, multistarts=10, biases=None, scales=None, seed=None)[source]#

Bases: object

Class that can be iterated to yield a set of random start points for a multistart NLP optimization problem (see csnlp.multistart.MultistartNlp and its subclasses).

Parameters:
pointsdict of (str, RandomStartPoint)

Dictionary containing the name of each variable, and how to generate random starting points for it (in the form of a RandomStartPoint object).

multistartsint, optional

The number of multiple start points. Default is 10.

biasesdict of (str, array_like), optional

Biases to add to the generated random points under the same name. If None, no bias is added.

scalesfloat, or array of floats

Scales to multiplty the generated random points with, under the same name. If None, no scale is multiplied.

seedNone, int, array_like of ints, SeedSequence, BitGenerator, Generator

RNG seed.

Methods

Examples using csnlp.multistart.RandomStartPoints#

A simple problem with multiple minima

A simple problem with multiple minima