Letteratura scientifica selezionata sul tema "Résilience aux erreurs"
Cita una fonte nei formati APA, MLA, Chicago, Harvard e in molti altri stili
Consulta la lista di attuali articoli, libri, tesi, atti di convegni e altre fonti scientifiche attinenti al tema "Résilience aux erreurs".
Accanto a ogni fonte nell'elenco di riferimenti c'è un pulsante "Aggiungi alla bibliografia". Premilo e genereremo automaticamente la citazione bibliografica dell'opera scelta nello stile citazionale di cui hai bisogno: APA, MLA, Harvard, Chicago, Vancouver ecc.
Puoi anche scaricare il testo completo della pubblicazione scientifica nel formato .pdf e leggere online l'abstract (il sommario) dell'opera se è presente nei metadati.
Articoli di riviste sul tema "Résilience aux erreurs"
Lichy, Jessica, e Derek Watson. "Managing uncertainty in food supply chains – The case of British food manufacturing". Question(s) de management 45, n. 4 (26 luglio 2023): 181–95. http://dx.doi.org/10.3917/qdm.225.0181.
Testo completoEl-Khodary, Mohammed. "ECONOMIC GROWTH DETERMINANTS IN MOROCCO: SHORT AND LONG-TERM ANALYSIS ON THE IMPACT OF FOREIGN AND DOMESTIC DIRECT INVESTMENT AND EXPORTS / DÉTERMINANTS DE LA CROISSANCE ÉCONOMIQUE AU MAROC : ANALYSE A COURT ET LONG TERME DE L'IMPACT DE L'INVESTISSEMENT DIRECT ET DES EXPORTATIONS". European Journal of Economic and Financial Research 8, n. 2 (19 aprile 2024). http://dx.doi.org/10.46827/ejefr.v8i2.1687.
Testo completoTesi sul tema "Résilience aux erreurs"
Piri, Ali. "Design space exploration for accuracy-aware computing". Electronic Thesis or Diss., Ecully, Ecole centrale de Lyon, 2024. http://www.theses.fr/2024ECDL0052.
Testo completoThe increasing complexity of large-scale applications, such as scientific computing and data analytics, has driven growing demands for computational resources. Traditional methods are becoming insufficient to handle the surging data volume. This thesis addresses the gap between data growth and processor advancements by introducing strategies for improving computational efficiency without sacrificing performance. A key focus is Approximate Computing (AxC), which leverages the inherent error tolerance in many applications—such as multimedia processing and machine learning—to balance performance with acceptable error levels. We conducted a comprehensive review of circuit-level functional approximations in various multipliers, aiming to deepen our understanding of the techniques used. We also discussed key accuracy metrics for evaluating the performance and reliability of these operators at the application and system levels. In addition, we reviewed automatic methods from the literature that support the broader adoption of approximate circuits, highlighting the need for a systematic approach to their design. The key contribution of the thesis is the introduction of input-aware approximate computing, which advances the design of efficient systems by optimizing performance and energy consumption while maintaining acceptable accuracy. The goal was to develop methods that are broadly applicable but specifically optimized for certain workloads. Initially, manual identification of circuit sections for approximation highlighted scalability challenges and the need for specialized expertise. To address this, an automated design approach was implemented, optimizing circuits based on specific input distributions. Tested with the FIT filter, this approach consistently improved accuracy and energy efficiency compared to traditional methods. Input-aware approximation is especially valuable for applications with predictable input patterns, such as embedded systems, machine learning inference, and real-time signal processing. This thesis also explores the relationship between approximate computing and the resilience of deep neural networks (DNNs), which are crucial in many applications. It analyzes how DNNs, especially those implemented on custom systolic arrays, respond to faults from approximate multipliers, using a Fault Injector Framework for efficient simulation. The study found that while DNNs are inherently fault-tolerant, approximations can significantly impact accuracy. A per-bit analysis revealed that critical bits, like the most significant bit (MSB), are particularly affected. Additionally, the trade-off between energy efficiency and resilience was highlighted, showing that energy savings from approximate computing must be balanced with the potential for performance degradation
Dkhili, Nouha. "A model-based predictive control strategy for low-voltage power distribution grids with prolific distributed generation in Occitania". Thesis, Perpignan, 2020. https://theses-public.univ-perp.fr/2020PERP0037.pdf.
Testo completoThe term ''smart grid'' refers to a modern power grid that successfully integrates prolific distributed generation with end loads and efficiently reroutes power flows to balance supply and demand in real time with respect to stability, quality, and safety constraints. It relies on improved observability and advanced control techniques, and offers the possibility of advanced demand side management.In the context of the Smart Occitania project, which aims to study the feasibility of the smart grid concept for rural and suburban power distribution grids, this work proposes a model-based predictive control strategy based on flexible asset management (herein a biogas plant and a water tower) that aims to balance power supply and demand within the power grid while maintaining voltage levels within prescribed margins. The control scheme incorporates intraday forecasts of various stochastic quantities that impact the system, procured through Gaussian process regression.The main contribution of this thesis is twofold: the predictive controller's optimisation problem is formulated in such a way that the ON/OFF of the water tower is handled without recourse to mixed-integer nonlinear programming or relaxation, and the confidence intervals provided by the forecast module are utilised to minimize voltage overshooting due to forecasting errors.The results illustrate the promise of a predictive controller relying on renewable-energy-based flexible assets to reduce the gap between power supply and demand, while upholding the power grid's voltage constraints
Capitoli di libri sul tema "Résilience aux erreurs"
CLAIRAY, Philippe. "Epidémies et pandémies : quelles leçons de l’histoire?" In Les épidémies au prisme des SHS, 115–26. Editions des archives contemporaines, 2022. http://dx.doi.org/10.17184/eac.5997.
Testo completo