Programação imunológica gramatical para inferência automática de modelos e projeto ótimo de estruturas

• Ano de defesa: 2012
• Autor(a) principal: Bernardino, Heder Soares
• Orientador(a): Barbosa, Helio José Corrêa
• Banca de defesa: Loula, Abimael Fernando Dourado , Dardenne, Laurent Emmanuel , Ebecken, Nelson Francisco Favilla , Takahashi, Ricardo Hiroshi Caldeira
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Laboratório Nacional de Computação Científica
• Programa de Pós-Graduação: Programa de Pós-Graduação em Modelagem Computacional
• Departamento: Serviço de Análise e Apoio a Formação de Recursos Humanos
• País: BR
• Palavras-chave em Português: Algorítmos genéticos; Projeto estrutural
• Palavras-chave em Inglês: Genetic algorithms; Structural design
• Área do conhecimento CNPq: CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://tede.lncc.br/handle/tede/154

Much effort has been made in order to automate the process of scientific knowledge discovery. Although some of the steps can be automated, the replacement of the specialist for a computer system in activities that require creativity remains a great challenge. In the field of computational intelligence the genetic programming techniques, and grammatical evolution in particular, appear to be adequate to these tasks. The use of formal grammars restricts the syntax of possible artifacts, making it possible to add bias in the search and generate more understandable solutions. Moreover, grammatical evolution establishes a clear distinction between the search space and the solution space, offering more flexibility. Thus, we propose here the grammar-based immune programming, a technique for evolving programs which combines a search engine inspired by clonal selection theory with the representation via grammatical evolution. Additionally, we developed a new procedure for decoding candidate solution that always generates a valid program. Finally, the technique is applied to symbolic regression and model inference in the form of ordinary differential equations. Its applicability in mechanical engineering is also exemplified in supporting the modeling of strains in a deformed pipe, and assisting the creation of optimal structural design