Uma arquitetura utilizando algoritmo genético interativo e aprendizado de máquina aplicado ao problema do próximo release

• Ano de defesa: 2015
• Autor(a) principal: Araújo, Allysson Allex de Paula
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Problema do próximo release; Algoritmo genético interativo; Otimização interativa; Aprendizado de máquina; Engenharia de Software baseada em busca
• Link de acesso: http://www.repositorio.ufc.br/handle/riufc/44988

The Next Release Problem consists in selecting which requirements will be implemented in the next software release. For many Search Based Software Engineering approaches to the Next Release Problem, there is still a lack of ability to efficiently include the human opinion and its peculiarities in the search process. Thus, in this work it is proposed an architecture to solve the Next Release Problem where the human preferences can be incorporated and, through a learning model, is able to minimize the problem of human fatigue. Experimental results are able to show that an Interactive Genetic Algorithm can successfully incorporate the user preferences in the final solution with a small loss in objectives terms.