Archivio Istituzionale della Ricerca - Università degli Studi di Pavia
Not a member yet
135341 research outputs found
Sort by
Imitation learning-driven approximation of stochastic control models
No full textPredictive control is a widely adopted methodology in numerous industries to manage multi-variable control problems under constraints. Traditional predictive control methods, however, assume complete model knowledge and deterministic behavior, which is often unrealistic in practical applications due to uncertainties and disturbances. Stochastic predictive control addresses these limitations by incorporating probabilistic models to enhance robustness in uncertain environments, but this approach comes at the cost of significantly increased computational complexity, making real-time implementation challenging. This paper proposes a method based on imitation learning to approximate the solution of stochastic predictive control, significantly reducing computational burden while maintaining predictive capabilities and robustness. The effectiveness of the proposed method is first demonstrated on the cart-pole stabilization problem, followed by its application to optimal lithium-ion battery charging. Results from both case studies underscore the method’s robust approximation capabilities and substantial computational cost reduction, underscoring its potential for real-time applications in diverse domains, including robotics, energy management, and autonomous systems
Phase field modeling of fracture nucleation and propagation: A theoretical and computational study
No full textThis thesis presents both theoretical results - most notably, a rigorous proof that steady-state phase-field evolutions satisfy Griffith’s criterion and Gamma-convergence analyses for new higher-order and cohesive models - and practical advances in the numerical simulation of complex fracture processes.
Chapter 1 provides an introduction and an historical and conceptual overview of fracture mechanics, detailing sharp-crack models, diffuse phase-field formulations and their relationship. It also discusses the inherent numerical challenges of both frameworks, emphasizing the need for robust algorithms that ensure convergence and physically meaningful evolutions.
In Chapter 2, a class of phase-field energies incorporating both volumetric–deviatoric and spectral decompositions is considered. The time-continuous limits of the resulting evolutions are analyzed, proving that in the steady-state regime
they satisfy a phase- field version of Griffith’s criterion in terms of toughness and a suitably defined energy release rate.
This analysis - valid independently of the adopted incremental scheme - clarifies the deep connection between the phase-field and sharp-crack approaches and is complemented by numerical simulations that compare their respective evolutions.
Within the phase-field framework, several choices exist in the literature for both the elastic and fracture energies. Along this line, Chapter 3 introduces a novel fourth-order AT1 phase-field model.
For the proposed functional, a rigorous Gamma-convergence result is established, and numerical simulations confirm its acuracy and efficiency.
AT1 models naturally account for crack nucleation, however their effective strength surface in stress space is elliptic, failing to reproduce the strong asymmetry between tensile and compressive responses observed experimentally. Several extensions based on energy decompositions have been proposed to address this limitation;
nevertheless, the flexibility of such models remains limited, reflecting an intrinsic feature of Griffith’s theory, which lacks a notion of material strength.
Historically, the conceptual gap between strength-based and energy-based approaches was bridged by cohesive zone models and in Chapter 4 a cohesive phase-field functional recently proposed by Vicentini et al. is studied. This model enables explicit control over the shape of the strength surface while decoupling strength from the regularization parameter. A Gamma-convergence analysis toward a sharp cohesive fracture model is presented in both one- and two-dimensional (antiplane) settings, using a finite element discrete formulation and exploiting the strong localization of the damage variable.
Numerical simulations explore the sensitivity of the model to mesh anisotropy, offering insight into both its theoretical robustness and its practical implementation.
Finally, Chapter 5 introduces an alternative route toward enhanced flexibility in defining strength surfaces, grounded in the theory of generalized standard materials. The dissipation potential is allowed to depend explicitly on strain. This formulation permits direct control over the shape of the domains while maintaining a fully variational structure.This thesis presents both theoretical results - most notably, a rigorous proof that steady-state phase-field evolutions satisfy Griffith’s criterion and Gamma-convergence analyses for new higher-order and cohesive models - and practical advances in the numerical simulation of complex fracture processes.
Chapter 1 provides an introduction and an historical and conceptual overview of fracture mechanics, detailing sharp-crack models, diffuse phase-field formulations and their relationship. It also discusses the inherent numerical challenges of both frameworks, emphasizing the need for robust algorithms that ensure convergence and physically meaningful evolutions.
In Chapter 2, a class of phase-field energies incorporating both volumetric–deviatoric and spectral decompositions is considered. The time-continuous limits of the resulting evolutions are analyzed, proving that in the steady-state regime
they satisfy a phase- field version of Griffith’s criterion in terms of toughness and a suitably defined energy release rate.
This analysis - valid independently of the adopted incremental scheme - clarifies the deep connection between the phase-field and sharp-crack approaches and is complemented by numerical simulations that compare their respective evolutions.
Within the phase-field framework, several choices exist in the literature for both the elastic and fracture energies. Along this line, Chapter 3 introduces a novel fourth-order AT1 phase-field model.
For the proposed functional, a rigorous Gamma-convergence result is established, and numerical simulations confirm its acuracy and efficiency.
AT1 models naturally account for crack nucleation, however their effective strength surface in stress space is elliptic, failing to reproduce the strong asymmetry between tensile and compressive responses observed experimentally. Several extensions based on energy decompositions have been proposed to address this limitation;
nevertheless, the flexibility of such models remains limited, reflecting an intrinsic feature of Griffith’s theory, which lacks a notion of material strength.
Historically, the conceptual gap between strength-based and energy-based approaches was bridged by cohesive zone models and in Chapter 4 a cohesive phase-field functional recently proposed by Vicentini et al. is studied. This model enables explicit control over the shape of the strength surface while decoupling strength from the regularization parameter. A Gamma-convergence analysis toward a sharp cohesive fracture model is presented in both one- and two-dimensional (antiplane) settings, using a finite element discrete formulation and exploiting the strong localization of the damage variable.
Numerical simulations explore the sensitivity of the model to mesh anisotropy, offering insight into both its theoretical robustness and its practical implementation.
Finally, Chapter 5 introduces an alternative route toward enhanced flexibility in defining strength surfaces, grounded in the theory of generalized standard materials. The dissipation potential is allowed to depend explicitly on strain. This formulation permits direct control over the shape of the domains while maintaining a fully variational structure
The Role of Machine Learning in LOS Reduction for Patients Affected by Lower Limb Fracture
No full textThe estimation of the Length of Stay (LOS) is a critical factor in clinical and managerial decision-making, helping healthcare professionals optimize hospital efficiency. For patients with orthopedic trauma, particularly those with lower limb fractures, LOS prediction becomes essential for resource planning and improving patient care. This study aims to analyze and predict LOS for patients with lower limb fractures admitted to the A.O.R.N. “Antonio Cardarelli” hospital in Naples. To achieve this, five neural network-based classifiers were implemented, and their performances were compared with those obtained in previous studies conducted by our research group, which employed well-established Artificial Intelligence (AI) models
The functional dynamics of FicD’s TPR domain are modulated by the interaction with ATP and BiP
No full textThe human Fic enzyme FicD plays an important role in regulating the Hsp70 homolog BiP in the endoplasmic reticulum: FicD reversibly modulates BiP’s activity through attaching an adenosine monophosphate to the substrate binding domain. This reduces BiP’s chaperone activity by shifting it into a conformation with reduced substrate affinity. Crystal structures of FicD in the apo, adenosine triphosphate (ATP)-bound, and BiP-bound states suggested significant conformational variability in the tetratricopeptide repeat (TPR) motifs. However, nothing is known about the underlying dynamics. In this study, we investigate the conformational dynamics of FicD’s TPR motifs using two-color, single-molecule Förster resonance energy transfer (smFRET) experiments. We demonstrate that the TPR motifs exhibit conformational dynamics between a TPR-out and a TPR-in conformation on timescales ranging from microseconds to milliseconds. In addition, we extend our investigation on multiple labeling positions within FicD, revealing how conformational dynamics vary depending on the location within the TPR motif. We quantify the motions with dynamic photon distribution analysis for the FRET constructs and generate an ensemble of structures for the different states consistent with the smFRET data using molecular dynamic simulations. We propose a conformational landscape model for FicD where the TPR-in/out states exist in equilibrium and the fraction of dynamic population is altered due to the presence of ATP and BiP. These results indicate that not only is FicD dynamic, but the dynamics are linked to the functionality and interactions of FicD with BiP
Entangled Food Histories. Methods and Sources for Exploring Foodways across Millennia
No full textEntangled Food Histories addresses a series of case studies that elucidate how food, drink and their associated activities have shaped and reflected transtemporal and transregional entanglements and disentanglements across millennia, from prehistory to the present. The volume brings together contributions from archaeologists, historians, and anthropologists to discuss the variety of sources and approaches that researchers can mobilize in the study of food history and archaeology. Ranging across Central America, Europe, Africa, and Asia, the volume underscores the significance of food as an analytical category to understand the cultural, economic, and political transformations that have shaped human societies across time and space
Quanv4EO: Empowering Earth Observation by Means of Quanvolutional Neural Networks
No full textA significant amount of remotely sensed data is generated daily by many Earth observation (EO) spaceborne and airborne sensors over different countries of our planet. Different applications use those data, such as natural hazard monitoring, global climate change, urban planning, and more. Many challenges are brought by the use of these big data in the context of remote sensing (RS) applications. In recent years, the employment of machine learning (ML) and deep learning (DL)-based algorithms has allowed a more efficient use of these data, but the issues in managing, processing, and efficiently exploiting them have even increased as classical computers have reached their limits. This article highlights a significant shift toward leveraging quantum computing (QC) techniques in processing large volumes of RS data. The proposed Quanv4EO framework introduces a quanvolution method for (pre)processing multidimensional EO data. Its effectiveness was first demonstrated on standard image classification datasets (MNIST and FashionMNIST), achieving accuracies of 99.84% and 96.81%, respectively, with a significantly reduced model size of 42 k parameters and 16 frozen qubits. Its capabilities were then checked on EO datasets, such as EuroSAT, with a mean accuracy of 96% using balanced iterative reducing and clustering using hierarchies (BIRCHs) clustering and 93% using automated DL (AutoDL), surpassing or matching state-of-the-art (SOTA) classical nonquantum models. Applying the framework to synthetic aperture radar (SAR) data, the QSPeckleFilter demonstrates notable improvements in speckle noise reduction, achieving a peak signal-to-noise ratio (PSNR) of 21.72 and a structural similarity index measure (SSIM) of 0.81, surpassing all tested classical counterparts. The proposed results underscore the potential of quantum-enhanced approaches in RS data analysis, paving the way for more efficient and effective solutions for wide geographical area EO data exploitation
Search for high-mass resonances in a final state comprising a gluon and two hadronically decaying W bosons in proton-proton collisions at = 13 TeV
No full textA search for high-mass resonances decaying into a gluon, g, and two W bosons is presented. A Kaluza-Klein gluon, gKK, decaying in cascade via a scalar radion R, gKK → gR → gWW, is considered. The final state studied consists of three large-radius jets, two of which contain the products of hadronically decaying W bosons, and the third one the hadronization products of the gluon. The analysis is performed using proton-proton collision data at s = 13 TeV collected by the CMS experiment at the CERN LHC during 2016–2018, corresponding to an integrated luminosity of 138 fb−1. The masses of the gKK and R candidates are reconstructed as trijet and dijet masses, respectively. These are used for event categorization and signal extraction. No excess of data events above the standard model background expectation is observed. Upper limits are set on the product of the gKK production cross section and its branching fraction via a radion R to gWW. This is the first analysis examining the resonant WW+jet signature and setting limits on the two resonance masses in an extended warped extra-dimensional model
Angular dependent measurement of electron-ion recombination in liquid argon for ionization calorimetry in the ICARUS liquid argon time projection chamber
No full textThis paper reports on a measurement of electron-ion recombination in liquid argon in the ICARUS liquid argon time projection chamber (LArTPC). A clear dependence of recombination on the angle of the ionizing particle track relative to the drift electric field is observed. An ellipsoid modified box (EMB) model of recombination describes the data across all measured angles. These measurements are used for the calorimetric energy scale calibration of the ICARUS TPC, which is also presented. The impact of the EMB model is studied on calorimetric particle identification, as well as muon and proton energy measurements. Accounting for the angular dependence in EMB recombination improves the accuracy and precision of these measurements
