Uma Abordagem para Seleção de Equipes Tecnicamente Qualificadas para Implementação de Projetos de Software Vinicius Souza dos Santos

• Autor(a) principal: Santos, Vinicius Souza dos
• Data de Publicação: 2014
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UFPB
• Texto Completo: https://repositorio.ufpb.br/jspui/handle/tede/6124
• topic: Informatica; Alocação de equipes; Software distribuido; Logica fuzzy; Representação de conhecimento; CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO

Due to the current needs and high demand for software products with more quality, various approaches for software development have been proposed. Among such approaches, Software Product Line has emerged as a promising approach for improving software quality and reducing costs and development time due to its high capacity for promoting software reuse. As another promising approach, Distributed Software Development has been adopted by organizations as a consequence of globalization, making possible to find skilled software professionals and more attractive costs in organizations spread throughout the world. By adopting such approaches together, domain experts can be found in several outsourcing or even insourcing software development teams around the world. However, even adopting such approaches to improve the software quality, is not possible to guarantee the software quality if software development teams do not have the necessary and specific knowledge to implement each module. So, a key question arises: which teams are more technically qualified to develop each module identified in a software project? Considering only ten software development teams and ten software modules, it is possible to identify 1010 combinations of modules and teams, in other words, 10.000.000.000 possible arrangements. Clearly, this is a complex process to be done by project managers based on their experience, since a large number of combinations can be identified with a relatively small number of teams and software modules. Thus any ad-hoc decision making can be complex, inefficient and error-prone. In order to provide support for recommending qualified global software teams in Software Product Line projects, this dissertation presents a decision support approach that identifies a set of technically qualified software development teams for each software module. To achieve such a goal, the proposed approach is based on technical requirements of software modules and technical capabilities of software development teams. The proposed approach is structured in four stages. The first stage is intended to produce an artifact that represents the technical requirements expected to develop each software module. Based on such requirements, the second stage is intended to represent the technical capabilities of all candidate software development teams. Then, the third stage aims to define a policy that guides the selection of such teams. Finally, the fourth stage maps the information collected in earlier stages into a fuzzy logic algorithm and performs the selection of technically qualified teams. As output, for each software module, a recommendation is generated that indicates a set of candidate teams that have the most appropriate technical skills to implement the software module