Multiobjective optimization with expensive functions. Survey on the state of the art

The multi-objective optimization is a complex process, even more when the functions that define the problems are not well conditioned or do not meet the minimum set requirements to ensure the convergence of classical algorithms, such as convexity, continuity and differentiability. Hence, the technic...

Descripción completa

Autores Principales: Calderón-Arce, Cindy, Alvarado-Moya, Pablo
Formato: Artículo
Idioma: Español
Publicado: Editorial Tecnológica de Costa Rica 2016
Materias:
Acceso en línea: https://revistas.tec.ac.cr/index.php/tec_marcha/article/view/2582
https://hdl.handle.net/2238/9014
Sumario: The multi-objective optimization is a complex process, even more when the functions that define the problems are not well conditioned or do not meet the minimum set requirements to ensure the convergence of classical algorithms, such as convexity, continuity and differentiability. Hence, the technical literature focuses on optimization techniques for problems defined by functions with specific characteristics, like high evaluation cost, non-convexity or non-differentiability. This article provides an overview of some of the prevailing techniques for problems with these kind of functions.