Otimização de redes de sensores sem fio: uma abordagem baseada em algoritmo genético para o posicionamento multiobjetivo em áreas urbanas

• Ano de defesa: 2025
• Autor(a) principal: Alves, Gabriel Sá Barreto
• Orientador(a): Jesus, Thiago Cerqueira de
• Banca de defesa: Pires, Matheus Giovanni, Peixoto, João Paulo Just, Loula, Angelo Conrado
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual de Feira de Santana
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciência da Computação
• Departamento: DEPARTAMENTO DE CIÊNCIAS EXATAS
• País: Brasil
• Palavras-chave em Português: Posicionamento de Sensores; Redes de Sensores sem Fio; Areas Urbanas; Algoritmo Genético; Otimização Multiobjetivo
• Palavras-chave em Inglês: Sensor Positioning; Wireless Sensor Networks; Urban Areas; Genetic Algorithm; Multi-objective Optimization
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://tede2.uefs.br:8080/handle/tede/1976

Technological advances have expanded the use of computing in various areas, highlighting Wireless Sensor Networks in the context of the Internet of Things. These networks enable continuous collection of structural and environmental data and rapid responses to critical events. However, the efficient placement of sensors is a central challenge, as inadequate arrangement can affect aspects such as coverage, energy consumption, and data reliability. This positioning problem is regarded as multi-objective due to the necessity of optimizing multiple criteria concurrently. In large-scale networks, combinatorial complexity precludes the application of traditional approaches in the pursuit of optimal solutions. In light of these findings, the present study puts forth and examines a novel positioning strategy for RSSFs comprising scalar and visual sensors. This strategy utilizes the Non-dominated Sorting Genetic Algorithm-II (NSGA-II) to optimize the criteria of coverage, sensing quality, and network connectivity. Among the experiments performed, the following should be emphasized: the definition of NSGA-II hyperparameters using Grid Search, the analysis of positioning in an ideal test scenario, the application of the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) for the selection of solutions on the Pareto Frontiers, and the analysis of the variability of the solutions found in one hundred scenarios with random network distribution. The results demonstrated the efficacy of the selected set of hyperparameters in identifying suitable solutions, meeting the established optimization criteria, and underscoring the efficiency of the proposed positioning process. Furthermore, the application of the TOPSIS method indicated that, for the present study, the selection of solutions can be made considering the same level of relevance for all criteria, since the prioritization did not show significant differences. Finally, the low variability observed in successive executions of NSGA-II demonstrated the stability and reproducibility of the proposed methodology in the search process for solutions. Consequently, this research can assist in the optimization of urban applications and facilitate the planning and adaptation of RSSFs in different outdoor environments. This contributes to the improvement of existing studies and provides guidelines for future research in the area.