2026-02-032026-02-032026-07-04BARATA, Claudio Matheus Modesto.Towards a robust b5g/6g transport network with self-adaptive network digital twin. Orientador: Aldebaro Barreto da Rocha Klautau Júnior. 2025. 82 f. Dissertação (Mestrado em Engenharia Elétrica) - Instituto de Tecnologia, Universidade Federal do Pará, Belém, 2025. Disponível em: https://repositorio.ufpa.br/handle/2011/17928. Acesso em:.https://repositorio.ufpa.br/handle/2011/17928The ability of the Network Digital Twin (NDT) to remain aware of changes in its physical counterpart, known as the Physical Twin (PTwin), is a fundamental condition to enable timely adaptation and synchronization, also referred to as twinning. In this way, considering a trans-port network, a key requirement is to handle unexpected traffic variability and dynamically adapt to maintain optimal performance in the associated virtual model, known as the Virtual Twin (VTwin). In this context, we propose a robust and self-adaptive implementation of a novel NDT arhitecture designed to provide accurate delay prediction for network flows, even under fluctuating traffic conditions. This architecture addresses an essential challenge, under-explored in the literature: improving the resilience of data-driven NDT platforms against traffic variability and improving synchronization between the VTwin, based on neural networks, and its physical counterpart. Therefore, the contributions of this dissertation rely on a relatively underexplored stage of the NDT lifecycle by focusing on the operational phase, where teleme-try modules are used to monitor incoming traffic and concept drift detection techniques guide retraining decisions aimed at updating and redeploying the VTwin when necessary. We vali-date our architecture across various emulated network topologies and diverse traffic patterns to demonstrate its effectiveness in preserving acceptable performance and predicting total per-flow delay under unexpected traffic variation, useful for maintaining reliable performance in appli-cations such as service-level agreement monitoring considered as a use case in this work. The results in all tested topologies, using the normalized mean square error as the evaluation metric, demonstrate that our proposed architecture, after a traffic concept drift, achieves a performance improvement in prediction of at least 56.7% compared to a configuration without digital twin synchronization.enAcesso AbertoAttribution-NonCommercial-NoDerivatives 4.0 Internationalhttp://creativecommons.org/licenses/by-nc-nd/4.0/Desvio de conceitoGerenciamento de ciclo de vidaRedes grafos neu- raisTwinningConcept driftGraph Neural Network (GNN)Lifecycle Management (LCM)TwinningDissertaçãoCNPQ::ENGENHARIAS::ENGENHARIA ELETRICA::TELECOMUNICACOESPROCESSAMENTO DE SINAISTELECOMUNICAÇÕES