Design and classification of dynamic multi-objective optimization problems

TL;DR

This paper introduces a formal model and classification for dynamic multi-objective optimization problems, focusing on time-dependent components affecting parameters, objectives, system states, and environment changes.

Contribution

It provides a novel formal framework and classification scheme for dynamic multi-objective optimization problems based on their time-dependent components.

Findings

Four main classes of dynamic components are identified.

Examples illustrate each class of dynamic component.

The model aids in understanding and designing algorithms for dynamic problems.

Abstract

In this work we provide a formal model for the different time-dependent components that can appear in dynamic multi-objective optimization problems, along with a classification of these components. Four main classes are identified, corresponding to the influence of the parameters, objective functions, previous states of the dynamic system and, last, environment changes, which in turn lead to online optimization problems. For illustration purposes, examples are provided for each class identified - by no means standing as the most representative ones or exhaustive in scope.