Sequences of Pareto reflections#

What are sequences of Pareto reflections? Sequences of Pareto reflections are a Pythonic implementation of a mathematical (possibly finite) sequence of Pareto reflections.

Why using sequences of Pareto reflections? With sequences of Pareto reflections we can target different (possibly excluding) properties of Pareto points within one MOO run. For example, we have Pareto reflections which target a certain corner of the Pareto front. By constructing a sequence of Pareto reflections using different Pareto reflections which target different corners we can construct a sequence which targets all corners.

How do we use sequences of Pareto reflections? This is done by calling the apply_to_sequence method of some MOO algorithm (implemented in the ParefMOO interface) to the sequence of Pareto reflections.

Currently, Paref includes implementations of the following sequences of Pareto reflections (illustrated by their corresponding property):

Property

Graphic

Sequence

Supported target space dimension

Note

Code

Filling gaps of Pareto front

Fill Gaps

FillGapsOfParetoFrontSequence

All

Being the edge points of the Pareto front

Fill Gaps

FindEdgePointsSequence

All

Repeating a (list of) Pareto reflections (generic)

RepeatingSequence

All

Generic sequence

Repeating a single Pareto reflection until a stopping criterion is met (generic)

NextWhenStoppingCriteriaMet

All

Generic Sequence