paref.pareto_reflection_sequences.generic.next_when_stopping_criteria_met#

Classes

NextWhenStoppingCriteriaMet(**kwargs)

Define a sequence by moving on to the next Pareto reflection if convergence is reached

class paref.pareto_reflection_sequences.generic.next_when_stopping_criteria_met.NextWhenStoppingCriteriaMet(**kwargs)[source]#

Bases: SequenceParetoReflections

Define a sequence by moving on to the next Pareto reflection if convergence is reached

When to use#

This sequence should be used if you want to repeat a single Pareto reflection until the Pareto point it is looking for is sufficiently close approximated (to be specified in a stopping criteria).

What it does#

The sequence simply loops through given Pareto reflections until a defined stopping criteria is met. If the end is reached, this sequence returns None, indicating the end of the sequence.

Examples

# TBA: add Initialize list of Pareto reflections

>>> import numpy as np
>>> from paref.pareto_reflections.restrict_by_point import RestrictByPoint
>>> from paref.pareto_reflection_sequences.generic.repeating_sequence import RepeatingSequence
>>> from paref.moo_algorithms.stopping_criteria.max_iterations_reached import MaxIterationsReached

Initialze stopping criteria

>>> stopping_criteria = MaxIterationsReached(max_iterations=1)

Initialize Pareto reflection to be repeated

>>> pareto_reflecting_functions = [RestrictByPoint(nadir=np.ones(1),restricting_point=np.ones(1))]

Initialize repeating sequence

>>> sequence = RepeatingSequence(pareto_reflections=pareto_reflecting_functions)

The repeating sequence returns the given Pareto reflection in each step of iteration until the stopping criteria is met

>>> sequence.next().__class__.__name__
RestrictByPoint

>>> sequence.next()
None

Initialize storage for Pareto reflections

next(**kwargs)#: Store Pareto reflections