http://hdl.handle.net/10071/15079 2026-04-16T17:44:42Z 2026-04-16T17:44:42Z Barros, J. Turner, C. http://hdl.handle.net/10071/36913 2026-04-15T14:57:44Z 2026-01-01T00:00:00Z

Título próprio: Network algorithm to model automotive supply chain structure Autoria: Barros, J.; Turner, C. Resumo: A network algorithm that models the structure of automotive supply chains, compiled from a proprietary database, is presented. An initial structural analysis was conducted using key performance indicators, including average path length, clustering coefficient, and degree distribution, to assess network configurations. The networks were then partitioned into subnetworks, with an emphasis on reflecting the operational dynamics of supply chain activities. Regression analysis was applied to each subnetwork, using the number of vertices as the independent variable, to develop an algorithm for generating synthetic networks. These synthetic constructs serve as benchmarks for the automotive sector and have shown a strong average correlation (0.94) with the structure of actual supply networks. This methodological contribution provides tools for analysing and optimising supply chain structures that underpin automotive engineering and manufacturing, ensuring robustness and efficiency in vehicle production systems. The prevalence of tree-like structures within supply networks challenge conventional beliefs regarding the complexity of automotive supply chains and prompts further investigation into the determinants of their resilience.

2026-01-01T00:00:00Z Caio, P. Silva, C. J. http://hdl.handle.net/10071/36813 2026-04-07T09:03:44Z 2025-01-01T00:00:00Z

Título próprio: Application of indirect methods to optimal control problems in epidemiology Autoria: Caio, P.; Silva, C. J. Editor: Aguiar, Antonio Pedro; Malonek, Paula Rocha; Pinto, Vítor Hugo; Fontes, Fernando A. C. C.; Chertovskih, Roman Resumo: Currently most of the numerical resolution of optimal control problems is done using direct methods where the increase accuracy of indirect methods is overshadowed by the necessary analytical derivation required beforehand. With recent developments from the control-toolbox ecosystem team the application of indirect methods as become more streamline enabling a wider range of problems to be solved, like, for example, optimal control problems applied to the transmission of infectious diseases. In this work, we aim to extend the application of indirect methods to optimal control problems applied to epidemiological models, using the control-toolbox

2025-01-01T00:00:00Z Yamaguchi, C. Stefenon, S. de Paz Santana, J. F. Leithardt, V. http://hdl.handle.net/10071/36783 2026-04-01T09:10:30Z 2025-01-01T00:00:00Z

Título próprio: Are graph neural networks better than standard classifiers? Autoria: Yamaguchi, C.; Stefenon, S.; de Paz Santana, J. F.; Leithardt, V. Editor: Iglesia, Daniel H. de la; de Paz Santana, Juan F.; López Rivero, Alfonso J. Resumo: Graph neural networks (GNNs) are becoming very popular these days due to their ability to perform classification and prediction depending on node connections. Since features of samples belonging to the same class can be related, graph-based models may perform classification better than other classifiers. The big challenge for this evaluation is to know if there is a sufficiently adequate relationship between the connections of the nodes to justify the use of these models, since connections to unrelated classes can reduce the capacity of these models. This paper proposes a thorough comparative evaluation between graph models and other well-established classifiers to assess the extent to which GNNs may be superior. This will be done by evaluating the relationships between the probability of connections between nodes and changing database features. The evaluation is performed using synthetic data, which is a task that can be evaluated in future work. The results show that when there are connections of classes different from the node under evaluation, the GNNs lose their advantages over other classifiers.

2025-01-01T00:00:00Z Costa, P. Noetzold, D. García Ovejero, R. Martín Esteban, R. Leithardt, V. http://hdl.handle.net/10071/36782 2026-04-01T08:34:57Z 2025-01-01T00:00:00Z

Título próprio: Prisec II: A comprehensive model for IoT security Autoria: Costa, P.; Noetzold, D.; García Ovejero, R.; Martín Esteban, R.; Leithardt, V. Editor: Iglesia, Daniel H. de la; Paz Santana, Juan F. de; López Rivero, Alfonso J. Resumo: This study examines data security and efficiency in interconnected IoT devices, focusing on selecting cryptographic algorithms to improve secure data transmission in 5G networks. The proposed model introduces four security levels, each applying different encryption strategies to balance security and performance. Cloud computing is integrated to mitigate computational limitations in IoT devices, optimizing encryption and decryption processes. The study evaluates multiple cryptographic algorithms, analyzing encryption and decryption times, packet throughput, and memory usage. Experimental results demonstrate that the model enables efficient encryption while maintaining security, with AES-256 and XChaCha20 showing stable performance across different packet sizes. The cloud-based implementation improves resource distribution and reduces processing delays. This work contributes a structured cryptographic model adaptable to varying security needs and a performance evaluation of different cryptographic approaches in cloud-integrated environments.

2025-01-01T00:00:00Z