HAL: Hyper Article en Ligne
Not a member yet
3159010 research outputs found
Sort by
Machine learning-driven solutions for sustainable and dynamic flexible job shop scheduling under worker absences and renewable energy variability
No full textInternational audienceThis paper adresses the Dynamic Sustainable Flexible Job Shop Scheduling Problem (DSFJSSP) by going beyond the traditionally emphasized economic dimension-such as makespan, flow time, or resource utilization-to include human and environmental factors, along with their related disruptions. Specifically, it considers human-related constraints such as workers' skills and ergonomic risks, as well as environmental aspects like carbon emissions from operations. Additionally, the study investigates the impact of worker absences and variability in renewable energy availability. To solve this problem, a multi-objective non-linear integer programming model is developed and an improved Non-dominated Sorting Genetic Algorithm III (INSGA-III) is employed et generate the initial scheduling solutions. Three Machine Learning (ML)-based approaches-Q-Learning, Deep Learning, and Deep Q-Learning-are used to determine the most effective rescheduling strategy in response to disruptions. Results show that partial rescheduling maintains a good balance across all objectives and a close adherence to the initial schedule. The right shift strategy is efficient for minor disruptions, while total rescheduling, though potentially effective, is time-consuming and can significantly deviate from the original schedule. The comparison of the considered ML methods confirms that the DQL offers the best adaptability and solution quality for selecting optimal rescheduling strategies. These results underscore the importance of adaptive scheduling in enhancing the resilience and sustainability of dynamic flexible job shop systems
Community challenge towards consensus on characterization of biological tissue: C4Bio’s first findings
No full textInternational audienceThis study investigates methodological variability across various expert laboratories worldwide, with regards to characterizing the mechanical properties of biological tissues. Two testing rounds were conducted on the specific use case of uniaxial tensile testing of porcine aorta. In the first round, 24 labs were invited to apply their established methods to assess inter-laboratory variability. This revealed significant methodological diversity and associated variability in the stress–stretch results, underscoring the necessity for a standardized approach. In the second round, a consensus protocol was collaboratively developed and adopted by 19 labs in an attempt to minimize variability. This involved standardized sample preparation and uniformity in testing protocol, including the use of a common cutting and thickness measurement tool. Despite protocol harmonization, significant variability persisted across labs, which could not be solely attributed to inherent biological differences in tissue samples. These results illustrate the challenges in unifying testing methods across different research settings, underlining the necessity for further refinement of testing practices. Enhancing consistency in biomechanical experiments is pivotal when comparing results across studies, as well as when using the resulting material properties for in silico simulations in medical research
D2SFNet: Dual-domain spatial-frequency network for few-shot medical image segmentation
No full textInternational audienceFew-shot learning has attracted growing attention in medical image segmentation due to its ability to achieve accurate results with limited labeled data by leveraging prior knowledge. However, existing few-shot segmentation methods are typically restricted to a single dataset and rely solely on spatial features, which limits their ability to model fine-grained anatomical structures and overlooks the utility of related datasets commonly available in clinical practice. To address these challenges, we propose a dual-domain spatial-frequency network (D2SFNet) that integrates frequency-domain information and data from heterogeneous domains. Specifically, we design a dual-domain joint training strategy that incorporates both the target and auxiliary datasets into the learning process, where the target dataset provides task-specific information while the auxiliary dataset contributes generalizable representation cues. To mitigate domain shifts in dual-domain training and enhance intra-class consistency, we introduce a novel joint alignment mechanism combining intra-and inter-domain alignment. Moreover, we employ the discrete cosine transform to extract complementary frequency-domain representations, which are dynamically fused with spatial features through a novel dynamic spatial-frequency representation (DSFR) module. Extensive experiments on three widely used medical image segmentation benchmarks demonstrate that D2SFNet consistently outperforms existing state-of-the-art methods. The source code is available at https://github.com/qchi-code/D2SFNet
Comparison of non-Newtonian models in a bearing with a porous layer: Viscoelastic (Maxwell) versus micropolar (couple stress)
No full textThe tribology literature abounds with articles concerning the effects of non-Newtonian lubricant models on bearing behavior. Different categorizations are possible, but we look at a viscoelastic model, and a micropolar model. The bearing we consider contains a porous layer with flow properties described by the Darcy model. The two continuum models: the Upper Convected Maxwell model (UCM) and the Stokes Couple Stress model (CS) are based on entirely different underlying physical assumptions. In the UCM model, the fluid is characterized by a time scale, namely, the relaxation time. In the CS model, the fluid is characterized by a length scale, representing the microstructure size. However, when the thin film assumptions are applied, the governing equations of the two approaches look surprisingly similar. According to computed results, viscoelasticity tends to increase the pressure. This effect is far more pronounced in the steep inclination case. The couple stress length parameter likewise tends to increase pressure, at both moderate and steep inclination. In all cases the porosity tends to decrease the pressure, due to an effective softening of the confining surface
Cool-flame chemistry of the representative bio-hybrid fuel 1,3-dioxane
No full textInternational audienceThe low-temperature oxidation of 1,3-dioxane was systematically investigated in two atmospheric pressure jetstirred reactors (JSRs) at temperatures ranging from 450 to 850 K and equivalence ratios ranging from 0.25 to 0.5. A suite of oxidation intermediates, including carbonyl compounds, conjugated olefins, cyclic ethers, and reactive hydroperoxides, were identified and quantified using synchrotron vacuum ultraviolet photoionization mass spectrometry (SVUV-PIMS) and gas chromatography (GC). The experiments reveal strong low-temperature reactivity and a pronounced negative temperature coefficient (NTC) behavior, that had not previously been reported for 1,3-dioxane to such an extent. A detailed kinetic model was developed to interpret the observed phenomena. Rate constants for hydrogen abstraction reactions by OH radicals, identified as the dominant fuel consumption pathways, are calculated using ab initio methods. The model also incorporates theoretically derived rate constants from the literature for key beta-scission ring-opening reactions and first-stage oxygen addition pathways of three 1,3-dioxanyl radicals. These inclusions improve the model's predictive capability and highlight the complex cool-flame chemistry associated with 1,3-dioxane. Model validation against experimental datasets from this study and literature-including JSR oxidation, flow reactor pyrolysis and oxidation, and ignition delay time (IDT) measurements-demonstrates good agreement. Mechanistic insights reveal that the ether group in the 1,3-dioxane ring facilitates hydrogen abstraction at the ortho positions (e.g., methylene bridge and ortho-CH2-) while suppressing abstraction at meta sites. These structural effects also influence intramolecular hydrogen shifts in ROO and OOQOOH radicals. Additionally, the presence of ring oxygen atoms weakens radical stabilization through inductive effects, promoting ring-opening reactions of R and QOOH species. Collectively, these factors contribute to the observed NTC behavior and the unique cool-flame characteristics of 1,3-dioxane
On the validity of the Moes iso-viscous - Rigid film thickness prediction, the influence of domain starvation and the parabolic geometry approximation
No full textThe validity of the Isoviscous Rigid film thickness prediction by Moes is analysed as well as the validity of the parabolic geometry approximating the spherical geometry
Low-temperature coprecipitation of γ- and ε-MnO2 nanomaterials: Role of counter-cation and addition rate
No full textInternational audienceManganese dioxide (MnO2) polymorphism governs its physicochemical properties and application performance in catalysis, electrochemistry, and magnetism. However, the role of Mn precursor counter-ions (KMnO4 vs. NaMnO4) and precursor addition kinetics in directing MnO2 crystal phase and defects architecture remains poorly understood. Here, we report a one-step aqueous coprecipitation route to selectively synthesize gamma-MnO2 or epsilon-MnO2 by controlling the counter-ion (Na+ vs. K+) of the MnO4- precursor (from NaMnO4 and KMnO4, respectively) and its addition rate. NaMnO4 yields phase-pure gamma-MnO2, whereas KMnO4 introduces K+ doping through charge transfer interactions, stabilizing epsilon-MnO2 through microtwinnings and De Wolff defects. These structural disorders modulate crystal growth pathways, producing anisotropic nanoflowers from entangled nanowires in epsilon-MnO2, in contrast to pseudo-hexagonal nanodisks from gamma-MnO2. The defect density dictates thermal stability and magnetic response, with epsilon-MnO2 exhibiting reduced antiferromagnetic coupling (Neel temperature of 10.2 K) compared to gamma-MnO2 (16.2 K). This work reveals a counter-ion-directed synthetic strategy to tune MnO2 polymorph, defect structure, and functional properties for targeted applications
Optimization of caper bud drying using the DT_LSBOOST model: A predictive approach to improve quality and efficiency
No full textInternational audienceCapparis spinosa L. buds undergo salting and drying to enhance their shelf life and organoleptic properties. This study evaluates the impact of four drying methods: oven drying (OD), vacuum drying (VD), freeze-drying (FD), and microwave drying (MD) on the physicochemical, antioxidant, and microbiological properties of dried caper buds. Salting reduced the initial moisture content from 508.50 % to 168.59 % (db), while drying further decreased it to approximately 9 %. Drying time varied significantly, with MD achieving the shortest duration (0.19-0.75h) and OD requiring the longest (reaching 49.66h). FD exhibited the highest energy consumption (60.77 kWh/kg), followed by VD, while OD and MD were the least energy-intensive (0.54-3.10 kWh/kg and 1.34-2.18 kWh/kg, respectively). FD preserved the most chlorophyll (193.63 mu g/g DW) and total phenolic content (28.98 mgGAE/g DW), whereas MD at 200 W resulted in the lowest TPC (9.88 mgGAE/g DW). FD samples also showed superior antioxidant activities in both ABTS and FRAP assays. In contrast, OD and MD increased browning and degraded quality attributes. Multivariate analyses (PCA and clustering) highlighted FD as optimal for preserving quality, while MD was the most detrimental. Microbiological analysis confirmed that dried capers met food safety standards. A predictive model using Decision Tree coupled with Least Squares Boosting (DT_LSBOOST) achieved exceptional accuracy (R = 0.9999, RMSE = 0.0564, ESP = 0.2028, MAE = 0.0305), providing a reliable tool for optimizing drying parameters. Overall, freeze-drying emerged as the best method to retain nutritional and bioactive properties of capers, and the developed predictive model offers an innovative approach to enhancing caper processing efficiency
EDWARD LORENZ’S CONTRIBUTION TO CHAOS THEORY
No full textInternational audienceThis chapter is devoted to the contribution of Edward Lorenz to chaos theory. It is based on some of his papers published in the 1960s and some other works that had some influence on his works