Les réseaux de données actuels sont des entités complexes qui opèrent dans des environnements dynamiques et hétérogènes. L'architecture et les protocoles de ces réseaux doivent faire face à plusieurs défis, notamment l'adaptation dynamique et la prise de décisions autonome en présence de plusieurs critères, souvent contradictoires, tels que le délai, le taux de perte, la gigue, l'énergie, etc. Cependant, les problèmes de décision multicritère ont généralement de multiples solutions. Ces problèmes sont résolus par des méthodes qui utilisent des paramètres dont les choix ont des conséquences difficiles à prévoir. De ce fait, la plupart des méthodes de décision multicritère proposées dans la littérature supposent la présence d'un décideur qui guide le processus de décision. Enfin, le choix des paramètres suppose souvent une interactivité avec le décideur, ce qui est difficile, voire impossible, à envisager dans un contexte autonome. Dans cette thèse, nous proposons une nouvelle méthode de décision multicritère adaptée aux systèmes autonomes en général et aux réseaux autonomes en particulier. La méthode de décision multicritère de type ``poupée russe'' que nous introduisons utilise un ensemble de boîtes de qualité englobantes, définies dans l'espace des critères, afin d'estimer une large gamme de fonctions d'utilité. D'une part, la méthode proposée s'adapte au caractère dynamique des réseaux autonomes, afin de maximiser la satisfaction des utilisateurs. D'autre part, elle utilise des paramètres qui sont soit directement déduits de faits objectifs, tels que des normes ou spécifications techniques, soit obtenus à l'aide d'une expérience de type MOS (Mean Opinion Score) au moyen d'une méthode de classification automatique. Nous avons testé les performances de la méthode de la poupée russe sur un cas pratique de routage dans les réseaux ad hoc sans fil. Les expérimentations ont montré que le routage réalisé avec la méthode de la poupée russe est toujours meilleur ou similaire à celui de la méthode de la somme pondérée qui est largement utilisée. Cela est dû à la capacité d'adaptation de la décision offerte par cette nouvelle méthode de décision multicritère / Today's data networks are complex entities that operate in dynamic and heterogeneous environments. The architecture and protocols of these networks have to face several challenges such as dynamic adaptation and autonomous decision-making in the presence of several, often conflicting, criteria such as delay, loss rate, jitter, energy, etc. However, multicriteria decision making problems usually have multiple solutions. These problems are solved by methods that use parameters whose choices have consequences difficult to predict. Thus, most multicriteria decision making methods proposed in the literature assume the presence of a decision maker who guides the decision process. Finally, the choice of parameter values often involves an interaction with the decision maker, which is difficult or impossible to do in an autonomous context. In this thesis, we propose a new multicriteria decision making method suitable for autonomous systems in general and autonomous networks in particular. The Russian doll like method we propose uses a set of nested quality boxes (like Russian dolls) defined in the criteria space, in order to approximate a wide range of utility functions. First, the proposed method adapts to the dynamic nature of autonomous networks in order to maximize user satisfaction. Second, it uses parameters that are directly deduced from objective facts such as technical standards or specifications, or obtained from a MOS (Mean Opinion Score) experiment using an automatic classification method. We tested the performance of the Russian doll like method in a case of routing in wireless ad hoc networks. Experiments have shown that routing done with the Russian doll like method is always better or similar to the one done by the weighted sum method which is widely used. This is due to the adaptation ability of the decision provided by this new multicriteria decision making method

[pt] Cenários de falhas em plataformas offshore constituem riscos ambientais, sociais e econômicos significativos devido à amplitude dos danos que podem ser provocados. Portanto, neste contexto, a adoção de boas práticas de gestão de manutenção é necessária, permitindo a continuidade operacional, através da integridade de ativos. No entanto, há ainda uma necessidade de estudos com uma abordagem holística para uma avaliação multinível de ativos, através de abordagens modernas que mitiguem as limitações da análise de modos e efeitos de falha (FMEA) tradicional. Desse modo, esta pesquisa propõe um framework para a análise FMEA que considera a subjetividade e as importâncias relativas dos critérios de risco, através de uma abordagem integrada que utiliza sistemas hierárquicos de inferência e modelos multicritério, como o Best-Worst Method (BWM) e o fuzzy VIsekriterijumsko KOmpromisno Rangiranje (VIKOR) para fornecer uma classificação estruturada de ativos, orientando o gestor na priorização de manutenção. Esse framework foi validado por meio de um estudo empírico em uma plataforma do tipo Floating Production Storage and Offloading (FPSO), que abordou a problemática da corrosão externa para gestão de manutenção, com base na coleta de dados triangulados. Portanto, a presente pesquisa traz contribuições teóricas ao preencher lacunas através da propositura e aplicação de metodologias integradas, e contribuições práticas ao fornecer aos gestores, responsáveis pela formulação de planos de manutenção, elementos objetivos para uma tomada de decisão mais eficaz, sob o paradigma da confiabilidade. Esse framework pode ser aplicado em outros domínios através do ajuste de critérios avaliativos. / [en] Failure scenarios in offshore platforms pose significant environmental, social, and economic risks due to the magnitude of potential damages they can cause. Therefore, in this context, the adoption of good maintenance management practices is necessary, enabling operational continuity through asset integrity. However, there is still a need for studies with a holistic approach for a multi-level assessment of assets, using modern approaches to mitigate the limitations of traditional Failure Mode and Effects Analysis (FMEA). Thus, this research proposes a framework for FMEA that considers the subjectivity and relative importance of risk criteria, through an integrated approach using hierarchical inference systems and multi-criteria models such as the Best-Worst Method (BWM) and fuzzy VIsekriterijumsko KOmpromisno Rangiranje (VIKOR) to provide a structured classification of assets, guiding managers in maintenance prioritization. This framework was validated through an empirical study on a Floating Production Storage and Offloading (FPSO) platform, addressing the issue of external corrosion for maintenance management, based on triangulated data collection. Therefore, this research brings theoretical contributions by filling gaps through the proposal and application of integrated methodologies, and practical contributions by providing managers responsible for formulating maintenance plans with objective elements for more effective decision-making under the reliability paradigm. This framework can be applied in other domains by adjusting evaluative criteria.