Hybrid Multi-Objective Evolutionary Algorithms with Collective Intelligence.

Cinalli, D.; Martí, L.; Sanchez-Pi, Nayat; Bicharra García A..C.

Abstract

Many real-world optimization problems can be formulated as multi-objective optimization problems (MOPs), in which two or more objective tunetions must be simultaneously optimized. This chapter covers some required formal definitions of multi-objective optimization and collective intelligence field. It outlines the usage of preferences and collective intelligence in Multi-Objective Evolutionary Algorithm (MOEAs), respectively. The chapter presents the new algorithms CI-NSGA-II and CI-SMS-EMOA based on interactive collective intelligence techniques. Since the beginning of 2000, the development of social network technologies and interactive online systems has promoted a broader understanding of the “intelligence” concept. The field known as collective intelligence is defined as the self-organized group intelligence arisen from participatory and collaboration actions of many individuals. Interactive genetic algorithms were successfully applied to get feedbacks of transitional results throughout the evolution process. MOEAs can handle intermediate non-dominated solutions to the decision maker and improve the search with a reference point or fitness function adjustments.

Más información

Editorial: CRC press
Fecha de publicación: 2018
Página de inicio: 48
Página final: 66
Idioma: Inglés
URL: https://www.addlabs.uff.br/Novo_Site_ADDLabs/images/documentos/publicacoes/publicacoes_pdf/capitulo_livros/2016/2016%20-%20Bio-Inspired.pdf
DOI:

https://doi.org/10.1201/9781315366845

Notas: .