-
A Comparative Analysis of Visual Odometry in Virtual and Real-World Railways Environments
Authors:
Gianluca D'Amico,
Mauro Marinoni,
Giorgio Buttazzo
Abstract:
Perception tasks play a crucial role in the development of automated operations and systems across multiple application fields. In the railway transportation domain, these tasks can improve the safety, reliability, and efficiency of various perations, including train localization, signal recognition, and track discrimination. However, collecting considerable and precisely labeled datasets for test…
▽ More
Perception tasks play a crucial role in the development of automated operations and systems across multiple application fields. In the railway transportation domain, these tasks can improve the safety, reliability, and efficiency of various perations, including train localization, signal recognition, and track discrimination. However, collecting considerable and precisely labeled datasets for testing such novel algorithms poses extreme challenges in the railway environment due to the severe restrictions in accessing the infrastructures and the practical difficulties associated with properly equip** trains with the required sensors, such as cameras and LiDARs. The remarkable innovations of graphic engine tools offer new solutions to craft realistic synthetic datasets. To illustrate the advantages of employing graphic simulation for early-stage testing of perception tasks in the railway domain, this paper presents a comparative analysis of the performance of a SLAM algorithm applied both in a virtual synthetic environment and a real-world scenario. The analysis leverages virtual railway environments created with the latest version of Unreal Engine, facilitating data collection and allowing the examination of challenging scenarios, including low-visibility, dangerous operational modes, and complex environments. The results highlight the feasibility and potentiality of graphic simulation to advance perception tasks in the railway domain.
△ Less
Submitted 25 March, 2024;
originally announced March 2024.
-
1-Lipschitz Layers Compared: Memory, Speed, and Certifiable Robustness
Authors:
Bernd Prach,
Fabio Brau,
Giorgio Buttazzo,
Christoph H. Lampert
Abstract:
The robustness of neural networks against input perturbations with bounded magnitude represents a serious concern in the deployment of deep learning models in safety-critical systems. Recently, the scientific community has focused on enhancing certifiable robustness guarantees by crafting 1-Lipschitz neural networks that leverage Lipschitz bounded dense and convolutional layers. Although different…
▽ More
The robustness of neural networks against input perturbations with bounded magnitude represents a serious concern in the deployment of deep learning models in safety-critical systems. Recently, the scientific community has focused on enhancing certifiable robustness guarantees by crafting 1-Lipschitz neural networks that leverage Lipschitz bounded dense and convolutional layers. Although different methods have been proposed in the literature to achieve this goal, understanding the performance of such methods is not straightforward, since different metrics can be relevant (e.g., training time, memory usage, accuracy, certifiable robustness) for different applications. For this reason, this work provides a thorough theoretical and empirical comparison between methods by evaluating them in terms of memory usage, speed, and certifiable robust accuracy. The paper also provides some guidelines and recommendations to support the user in selecting the methods that work best depending on the available resources. We provide code at https://github.com/berndprach/1LipschitzLayersCompared.
△ Less
Submitted 28 November, 2023;
originally announced November 2023.
-
Attention-Based Real-Time Defenses for Physical Adversarial Attacks in Vision Applications
Authors:
Giulio Rossolini,
Alessandro Biondi,
Giorgio Buttazzo
Abstract:
Deep neural networks exhibit excellent performance in computer vision tasks, but their vulnerability to real-world adversarial attacks, achieved through physical objects that can corrupt their predictions, raises serious security concerns for their application in safety-critical domains. Existing defense methods focus on single-frame analysis and are characterized by high computational costs that…
▽ More
Deep neural networks exhibit excellent performance in computer vision tasks, but their vulnerability to real-world adversarial attacks, achieved through physical objects that can corrupt their predictions, raises serious security concerns for their application in safety-critical domains. Existing defense methods focus on single-frame analysis and are characterized by high computational costs that limit their applicability in multi-frame scenarios, where real-time decisions are crucial.
To address this problem, this paper proposes an efficient attention-based defense mechanism that exploits adversarial channel-attention to quickly identify and track malicious objects in shallow network layers and mask their adversarial effects in a multi-frame setting. This work advances the state of the art by enhancing existing over-activation techniques for real-world adversarial attacks to make them usable in real-time applications. It also introduces an efficient multi-frame defense framework, validating its efficacy through extensive experiments aimed at evaluating both defense performance and computational cost.
△ Less
Submitted 18 November, 2023;
originally announced November 2023.
-
TrainSim: A Railway Simulation Framework for LiDAR and Camera Dataset Generation
Authors:
Gianluca D'Amico,
Mauro Marinoni,
Federico Nesti,
Giulio Rossolini,
Giorgio Buttazzo,
Salvatore Sabina,
Gianluigi Lauro
Abstract:
The railway industry is searching for new ways to automate a number of complex train functions, such as object detection, track discrimination, and accurate train positioning, which require the artificial perception of the railway environment through different types of sensors, including cameras, LiDARs, wheel encoders, and inertial measurement units. A promising approach for processing such senso…
▽ More
The railway industry is searching for new ways to automate a number of complex train functions, such as object detection, track discrimination, and accurate train positioning, which require the artificial perception of the railway environment through different types of sensors, including cameras, LiDARs, wheel encoders, and inertial measurement units. A promising approach for processing such sensory data is the use of deep learning models, which proved to achieve excellent performance in other application domains, as robotics and self-driving cars. However, testing new algorithms and solutions requires the availability of a large amount of labeled data, acquired in different scenarios and operating conditions, which are difficult to obtain in a real railway setting due to strict regulations and practical constraints in accessing the trackside infrastructure and equip** a train with the required sensors. To address such difficulties, this paper presents a visual simulation framework able to generate realistic railway scenarios in a virtual environment and automatically produce inertial data and labeled datasets from emulated LiDARs and cameras useful for training deep neural networks or testing innovative algorithms. A set of experimental results are reported to show the effectiveness of the proposed approach.
△ Less
Submitted 28 February, 2023;
originally announced February 2023.
-
On the regularity of optimal potentials in control problems governed by elliptic equations
Authors:
Giuseppe Buttazzo,
Juan Casado_Díaz,
Faustino Maestre
Abstract:
In this paper we consider optimal control problems where the control variable is a potential and the state equation is an elliptic partial differential equation of a Schrödinger type, governed by the Laplace operator. The cost functional involves the solution of the state equation and a penalization term for the control variable. While the existence of an optimal solution simply follows by the dir…
▽ More
In this paper we consider optimal control problems where the control variable is a potential and the state equation is an elliptic partial differential equation of a Schrödinger type, governed by the Laplace operator. The cost functional involves the solution of the state equation and a penalization term for the control variable. While the existence of an optimal solution simply follows by the direct methods of the calculus of variations, the regularity of the optimal potential is a difficult question and under the general assumptions we consider, no better regularity than the $BV$ one can be expected. This happens in particular for the cases in which a bang-bang solution occurs, where optimal potentials are characteristic functions of a domain. We prove the $BV$ regularity of optimal solutions through a regularity result for PDEs. Some numerical simulations show the behavior of optimal potentials in some particular cases.
△ Less
Submitted 5 February, 2023;
originally announced February 2023.
-
On the numerical approximation of Blaschke-Santaló diagrams using Centroidal Voronoi Tessellations
Authors:
Beniamin Bogosel,
Giuseppe Buttazzo,
Edouard Oudet
Abstract:
Identifying Blaschke-Santaló diagrams is an important topic that essentially consists in determining the image $Y=F(X)$ of a map $F:X\to{\mathbb{R}}^d$, where the dimension of the source space $X$ is much larger than the one of the target space. In some cases, that occur for instance in shape optimization problems, $X$ can even be a subset of an infinite-dimensional space. The usual Monte Carlo me…
▽ More
Identifying Blaschke-Santaló diagrams is an important topic that essentially consists in determining the image $Y=F(X)$ of a map $F:X\to{\mathbb{R}}^d$, where the dimension of the source space $X$ is much larger than the one of the target space. In some cases, that occur for instance in shape optimization problems, $X$ can even be a subset of an infinite-dimensional space. The usual Monte Carlo method, consisting in randomly choosing a number $N$ of points $x_1,\dots,x_N$ in $X$ and plotting them in the target space ${\mathbb{R}}^d$, produces in many cases areas in $Y$ of very high and very low concentration leading to a rather rough numerical identification of the image set. On the contrary, our goal is to choose the points $x_i$ in an appropriate way that produces a uniform distribution in the target space. In this way we may obtain a good representation of the image set $Y$ by a relatively small number $N$ of samples which is very useful when the dimension of the source space $X$ is large (or even infinite) and the evaluation of $F(x_i)$ is costly. Our method consists in a suitable use of {\it Centroidal Voronoi Tessellations} which provides efficient numerical results. Simulations for two and three dimensional examples are shown in the paper.
△ Less
Submitted 1 February, 2023;
originally announced February 2023.
-
Regularity of the optimal sets for a class of integral shape functionals
Authors:
Giuseppe Buttazzo,
Francesco Paolo Maiale,
Dario Mazzoleni,
Giorgio Tortone,
Bozhidar Velichkov
Abstract:
We prove {the first} regularity theorem for the free boundary of solutions to shape optimization problems involving integral functionals, for which the energy of a domain $Ω$ is obtained as the integral of a cost function $j(u,x)$ depending on the solution $u$ of a certain PDE problem on $Ω$. The main feature of these functionals is that the minimality of a domain $Ω$ cannot be translated into a v…
▽ More
We prove {the first} regularity theorem for the free boundary of solutions to shape optimization problems involving integral functionals, for which the energy of a domain $Ω$ is obtained as the integral of a cost function $j(u,x)$ depending on the solution $u$ of a certain PDE problem on $Ω$. The main feature of these functionals is that the minimality of a domain $Ω$ cannot be translated into a variational problem for a single (real or vector valued) state function.
In this paper we focus on the case of affine cost functions $j(u,x)=-g(x)u+Q(x)$, where $u$ is the solution of the PDE $-Δu=f$ with Dirichlet boundary conditions. We obtain the Lipschitz continuity and the non-degeneracy of the optimal $u$ from the inwards/outwards optimality of $Ω$ and then we use the stability of $Ω$ with respect to variations with smooth vector fields in order to study the blow-up limits of the state function $u$. By performing a triple consecutive blow-up, we prove the existence of blow-up sequences converging to homogeneous stable solution of the one-phase Bernoulli problem and according to the blow-up limits, we decompose $\partialΩ$ into a singular and a regular part. In order to estimate the Hausdorff dimension of the singular set of $\partialΩ$ we give a new formulation of the notion of stability for the one-phase problem, which is preserved under blow-up limits and allows to develop a dimension reduction principle. Finally, by combining a higher order Boundary Harnack principle and a viscosity approach, we prove $C^\infty$ regularity of the regular part of the free boundary when the data are smooth.
△ Less
Submitted 18 December, 2022;
originally announced December 2022.
-
On a reverse Kohler-Jobin inequality
Authors:
Luca Briani,
Giuseppe Buttazzo,
Serena Guarino Lo Bianco
Abstract:
We consider the shape optimization problems for the quantities $λ(Ω)T^q(Ω)$, where $Ω$ varies among open sets of $\mathbb{R}^d$ with a prescribed Lebesgue measure. While the characterization of the infimum is completely clear, the same does not happen for the maximization in the case $q>1$. We prove that for $q$ large enough a maximizing domain exists among quasi-open sets and that the ball is opt…
▽ More
We consider the shape optimization problems for the quantities $λ(Ω)T^q(Ω)$, where $Ω$ varies among open sets of $\mathbb{R}^d$ with a prescribed Lebesgue measure. While the characterization of the infimum is completely clear, the same does not happen for the maximization in the case $q>1$. We prove that for $q$ large enough a maximizing domain exists among quasi-open sets and that the ball is optimal among {\it nearly spherical domains}.
△ Less
Submitted 10 December, 2022;
originally announced December 2022.
-
On the continuity of the Continuous Steiner Symmetrization
Authors:
Giuseppe Buttazzo
Abstract:
Starting from the Brock's construction of Continuous Steiner Symmetrization of sets, the problem of modifying continuously a given domain up to obtain a ball, preserving its measure and with decreasing first eigenvalue of the Laplace operator, is considered. For a large class of cases it is shown this is possible, while the general question remains still open.
Starting from the Brock's construction of Continuous Steiner Symmetrization of sets, the problem of modifying continuously a given domain up to obtain a ball, preserving its measure and with decreasing first eigenvalue of the Laplace operator, is considered. For a large class of cases it is shown this is possible, while the general question remains still open.
△ Less
Submitted 9 November, 2022;
originally announced November 2022.
-
Robust-by-Design Classification via Unitary-Gradient Neural Networks
Authors:
Fabio Brau,
Giulio Rossolini,
Alessandro Biondi,
Giorgio Buttazzo
Abstract:
The use of neural networks in safety-critical systems requires safe and robust models, due to the existence of adversarial attacks. Knowing the minimal adversarial perturbation of any input x, or, equivalently, knowing the distance of x from the classification boundary, allows evaluating the classification robustness, providing certifiable predictions. Unfortunately, state-of-the-art techniques fo…
▽ More
The use of neural networks in safety-critical systems requires safe and robust models, due to the existence of adversarial attacks. Knowing the minimal adversarial perturbation of any input x, or, equivalently, knowing the distance of x from the classification boundary, allows evaluating the classification robustness, providing certifiable predictions. Unfortunately, state-of-the-art techniques for computing such a distance are computationally expensive and hence not suited for online applications. This work proposes a novel family of classifiers, namely Signed Distance Classifiers (SDCs), that, from a theoretical perspective, directly output the exact distance of x from the classification boundary, rather than a probability score (e.g., SoftMax). SDCs represent a family of robust-by-design classifiers. To practically address the theoretical requirements of a SDC, a novel network architecture named Unitary-Gradient Neural Network is presented. Experimental results show that the proposed architecture approximates a signed distance classifier, hence allowing an online certifiable classification of x at the cost of a single inference.
△ Less
Submitted 9 September, 2022;
originally announced September 2022.
-
Wasserstein interpolation with constraints and application to a parking problem
Authors:
Giuseppe Buttazzo,
Guillaume Carlier,
Katharina Eichinger
Abstract:
We consider optimal transport problems where the cost for transporting a given probability measure $μ_0$ to another one $μ_1$ consists of two parts: the first one measures the transportation from $μ_0$ to an intermediate (pivot) measure $μ$ to be determined (and subject to various constraints), and the second one measures the transportation from $μ$ to $μ_1$. This leads to Wasserstein interpolatio…
▽ More
We consider optimal transport problems where the cost for transporting a given probability measure $μ_0$ to another one $μ_1$ consists of two parts: the first one measures the transportation from $μ_0$ to an intermediate (pivot) measure $μ$ to be determined (and subject to various constraints), and the second one measures the transportation from $μ$ to $μ_1$. This leads to Wasserstein interpolation problems under constraints for which we establish various properties of the optimal pivot measures $μ$. Considering the more general situation where only some part of the mass uses the intermediate stop leads to a mathematical model for the optimal location of a parking region around a city. Numerical simulations, based on entropic regularization, are presented both for the optimal parking regions and for Wasserstein constrained interpolation problems.
△ Less
Submitted 28 July, 2022;
originally announced July 2022.
-
CARLA-GeAR: a Dataset Generator for a Systematic Evaluation of Adversarial Robustness of Vision Models
Authors:
Federico Nesti,
Giulio Rossolini,
Gianluca D'Amico,
Alessandro Biondi,
Giorgio Buttazzo
Abstract:
Adversarial examples represent a serious threat for deep neural networks in several application domains and a huge amount of work has been produced to investigate them and mitigate their effects. Nevertheless, no much work has been devoted to the generation of datasets specifically designed to evaluate the adversarial robustness of neural models. This paper presents CARLA-GeAR, a tool for the auto…
▽ More
Adversarial examples represent a serious threat for deep neural networks in several application domains and a huge amount of work has been produced to investigate them and mitigate their effects. Nevertheless, no much work has been devoted to the generation of datasets specifically designed to evaluate the adversarial robustness of neural models. This paper presents CARLA-GeAR, a tool for the automatic generation of photo-realistic synthetic datasets that can be used for a systematic evaluation of the adversarial robustness of neural models against physical adversarial patches, as well as for comparing the performance of different adversarial defense/detection methods. The tool is built on the CARLA simulator, using its Python API, and allows the generation of datasets for several vision tasks in the context of autonomous driving. The adversarial patches included in the generated datasets are attached to billboards or the back of a truck and are crafted by using state-of-the-art white-box attack strategies to maximize the prediction error of the model under test. Finally, the paper presents an experimental study to evaluate the performance of some defense methods against such attacks, showing how the datasets generated with CARLA-GeAR might be used in future work as a benchmark for adversarial defense in the real world. All the code and datasets used in this paper are available at http://carlagear.retis.santannapisa.it.
△ Less
Submitted 9 June, 2022;
originally announced June 2022.
-
Mass optimization problem with convex cost
Authors:
Giuseppe Buttazzo,
Maria Stella Gelli,
Danka Lučić
Abstract:
In this paper we consider a mass optimization problem in the case of scalar state function, where instead of imposing a constraint on the total mass of the competitors, we penalize the classical compliance by a convex functional defined on the space of measures. We obtain a characterization of optimal solutions to the problem through a suitable PDE. This generalizes the case considered in the lite…
▽ More
In this paper we consider a mass optimization problem in the case of scalar state function, where instead of imposing a constraint on the total mass of the competitors, we penalize the classical compliance by a convex functional defined on the space of measures. We obtain a characterization of optimal solutions to the problem through a suitable PDE. This generalizes the case considered in the literature of a linear cost and applies to the optimization of a conductor where very low and very high conductivities have both a high cost, and then the study of nonlinear models becomes relevant.
△ Less
Submitted 8 April, 2023; v1 submitted 11 April, 2022;
originally announced April 2022.
-
Defending From Physically-Realizable Adversarial Attacks Through Internal Over-Activation Analysis
Authors:
Giulio Rossolini,
Federico Nesti,
Fabio Brau,
Alessandro Biondi,
Giorgio Buttazzo
Abstract:
This work presents Z-Mask, a robust and effective strategy to improve the adversarial robustness of convolutional networks against physically-realizable adversarial attacks. The presented defense relies on specific Z-score analysis performed on the internal network features to detect and mask the pixels corresponding to adversarial objects in the input image. To this end, spatially contiguous acti…
▽ More
This work presents Z-Mask, a robust and effective strategy to improve the adversarial robustness of convolutional networks against physically-realizable adversarial attacks. The presented defense relies on specific Z-score analysis performed on the internal network features to detect and mask the pixels corresponding to adversarial objects in the input image. To this end, spatially contiguous activations are examined in shallow and deep layers to suggest potential adversarial regions. Such proposals are then aggregated through a multi-thresholding mechanism. The effectiveness of Z-Mask is evaluated with an extensive set of experiments carried out on models for both semantic segmentation and object detection. The evaluation is performed with both digital patches added to the input images and printed patches positioned in the real world. The obtained results confirm that Z-Mask outperforms the state-of-the-art methods in terms of both detection accuracy and overall performance of the networks under attack. Additional experiments showed that Z-Mask is also robust against possible defense-aware attacks.
△ Less
Submitted 15 September, 2022; v1 submitted 14 March, 2022;
originally announced March 2022.
-
On the Real-World Adversarial Robustness of Real-Time Semantic Segmentation Models for Autonomous Driving
Authors:
Giulio Rossolini,
Federico Nesti,
Gianluca D'Amico,
Saasha Nair,
Alessandro Biondi,
Giorgio Buttazzo
Abstract:
The existence of real-world adversarial examples (commonly in the form of patches) poses a serious threat for the use of deep learning models in safety-critical computer vision tasks such as visual perception in autonomous driving. This paper presents an extensive evaluation of the robustness of semantic segmentation models when attacked with different types of adversarial patches, including digit…
▽ More
The existence of real-world adversarial examples (commonly in the form of patches) poses a serious threat for the use of deep learning models in safety-critical computer vision tasks such as visual perception in autonomous driving. This paper presents an extensive evaluation of the robustness of semantic segmentation models when attacked with different types of adversarial patches, including digital, simulated, and physical ones. A novel loss function is proposed to improve the capabilities of attackers in inducing a misclassification of pixels. Also, a novel attack strategy is presented to improve the Expectation Over Transformation method for placing a patch in the scene. Finally, a state-of-the-art method for detecting adversarial patch is first extended to cope with semantic segmentation models, then improved to obtain real-time performance, and eventually evaluated in real-world scenarios. Experimental results reveal that, even though the adversarial effect is visible with both digital and real-world attacks, its impact is often spatially confined to areas of the image around the patch. This opens to further questions about the spatial robustness of real-time semantic segmentation models.
△ Less
Submitted 5 January, 2022;
originally announced January 2022.
-
On the Minimal Adversarial Perturbation for Deep Neural Networks with Provable Estimation Error
Authors:
Fabio Brau,
Giulio Rossolini,
Alessandro Biondi,
Giorgio Buttazzo
Abstract:
Although Deep Neural Networks (DNNs) have shown incredible performance in perceptive and control tasks, several trustworthy issues are still open. One of the most discussed topics is the existence of adversarial perturbations, which has opened an interesting research line on provable techniques capable of quantifying the robustness of a given input. In this regard, the Euclidean distance of the in…
▽ More
Although Deep Neural Networks (DNNs) have shown incredible performance in perceptive and control tasks, several trustworthy issues are still open. One of the most discussed topics is the existence of adversarial perturbations, which has opened an interesting research line on provable techniques capable of quantifying the robustness of a given input. In this regard, the Euclidean distance of the input from the classification boundary denotes a well-proved robustness assessment as the minimal affordable adversarial perturbation. Unfortunately, computing such a distance is highly complex due the non-convex nature of NNs. Despite several methods have been proposed to address this issue, to the best of our knowledge, no provable results have been presented to estimate and bound the error committed. This paper addresses this issue by proposing two lightweight strategies to find the minimal adversarial perturbation. Differently from the state-of-the-art, the proposed approach allows formulating an error estimation theory of the approximate distance with respect to the theoretical one. Finally, a substantial set of experiments is reported to evaluate the performance of the algorithms and support the theoretical findings. The obtained results show that the proposed strategies approximate the theoretical distance for samples close to the classification boundary, leading to provable robustness guarantees against any adversarial attacks.
△ Less
Submitted 4 January, 2022;
originally announced January 2022.
-
On a class of Cheeger inequalities
Authors:
Luca Briani,
Giuseppe Buttazzo,
Francesca Prinari
Abstract:
We study a general version of the Cheeger inequality by considering the shape functional $\mathcal{F}_{p,q}(Ω)=λ_p^{1/p}(Ω)/λ_q(Ω)^{1/q}$. The infimum and the supremum of $\mathcal{F}_{p,q}$ are studied in the class of all domains $Ω$ of $\mathbb{R}^d$ and in the subclass of convex domains. In the latter case the issue concerning the existence of an optimal domain for $\mathcal{F}_{p,q}$ is discus…
▽ More
We study a general version of the Cheeger inequality by considering the shape functional $\mathcal{F}_{p,q}(Ω)=λ_p^{1/p}(Ω)/λ_q(Ω)^{1/q}$. The infimum and the supremum of $\mathcal{F}_{p,q}$ are studied in the class of all domains $Ω$ of $\mathbb{R}^d$ and in the subclass of convex domains. In the latter case the issue concerning the existence of an optimal domain for $\mathcal{F}_{p,q}$ is discussed.
△ Less
Submitted 9 April, 2022; v1 submitted 25 November, 2021;
originally announced November 2021.
-
Evaluating the Robustness of Semantic Segmentation for Autonomous Driving against Real-World Adversarial Patch Attacks
Authors:
Federico Nesti,
Giulio Rossolini,
Saasha Nair,
Alessandro Biondi,
Giorgio Buttazzo
Abstract:
Deep learning and convolutional neural networks allow achieving impressive performance in computer vision tasks, such as object detection and semantic segmentation (SS). However, recent studies have shown evident weaknesses of such models against adversarial perturbations. In a real-world scenario instead, like autonomous driving, more attention should be devoted to real-world adversarial examples…
▽ More
Deep learning and convolutional neural networks allow achieving impressive performance in computer vision tasks, such as object detection and semantic segmentation (SS). However, recent studies have shown evident weaknesses of such models against adversarial perturbations. In a real-world scenario instead, like autonomous driving, more attention should be devoted to real-world adversarial examples (RWAEs), which are physical objects (e.g., billboards and printable patches) optimized to be adversarial to the entire perception pipeline. This paper presents an in-depth evaluation of the robustness of popular SS models by testing the effects of both digital and real-world adversarial patches. These patches are crafted with powerful attacks enriched with a novel loss function. Firstly, an investigation on the Cityscapes dataset is conducted by extending the Expectation Over Transformation (EOT) paradigm to cope with SS. Then, a novel attack optimization, called scene-specific attack, is proposed. Such an attack leverages the CARLA driving simulator to improve the transferability of the proposed EOT-based attack to a real 3D environment. Finally, a printed physical billboard containing an adversarial patch was tested in an outdoor driving scenario to assess the feasibility of the studied attacks in the real world. Exhaustive experiments revealed that the proposed attack formulations outperform previous work to craft both digital and real-world adversarial patches for SS. At the same time, the experimental results showed how these attacks are notably less effective in the real world, hence questioning the practical relevance of adversarial attacks to SS models for autonomous/assisted driving.
△ Less
Submitted 13 August, 2021;
originally announced August 2021.
-
Shape optimization problems in control form
Authors:
Giuseppe Buttazzo,
Francesco Paolo Maiale,
Bozhidar Velichkov
Abstract:
We consider a shape optimization problem written in the optimal control form: the governing operator is the $p$-Laplacian in the Euclidean space $\R^d$, the cost is of an integral type, and the control variable is the domain of the state equation. Conditions that guarantee the existence of an optimal domain will be discussed in various situations. It is proved that the optimal domains have a finit…
▽ More
We consider a shape optimization problem written in the optimal control form: the governing operator is the $p$-Laplacian in the Euclidean space $\R^d$, the cost is of an integral type, and the control variable is the domain of the state equation. Conditions that guarantee the existence of an optimal domain will be discussed in various situations. It is proved that the optimal domains have a finite perimeter and, under some suitable assumptions, that they are open sets. A crucial difference is between the case $p>d$, where the existence occurs under very mild conditions, and the case $p\le d$, where additional assumptions have to be made on the data.
△ Less
Submitted 25 June, 2021; v1 submitted 8 May, 2021;
originally announced May 2021.
-
Increasing the Confidence of Deep Neural Networks by Coverage Analysis
Authors:
Giulio Rossolini,
Alessandro Biondi,
Giorgio Buttazzo
Abstract:
The great performance of machine learning algorithms and deep neural networks in several perception and control tasks is pushing the industry to adopt such technologies in safety-critical applications, as autonomous robots and self-driving vehicles. At present, however, several issues need to be solved to make deep learning methods more trustworthy, predictable, safe, and secure against adversaria…
▽ More
The great performance of machine learning algorithms and deep neural networks in several perception and control tasks is pushing the industry to adopt such technologies in safety-critical applications, as autonomous robots and self-driving vehicles. At present, however, several issues need to be solved to make deep learning methods more trustworthy, predictable, safe, and secure against adversarial attacks. Although several methods have been proposed to improve the trustworthiness of deep neural networks, most of them are tailored for specific classes of adversarial examples, hence failing to detect other corner cases or unsafe inputs that heavily deviate from the training samples. This paper presents a lightweight monitoring architecture based on coverage paradigms to enhance the model robustness against different unsafe inputs. In particular, four coverage analysis methods are proposed and tested in the architecture for evaluating multiple detection logics. Experimental results show that the proposed approach is effective in detecting both powerful adversarial examples and out-of-distribution inputs, introducing limited extra-execution time and memory requirements.
△ Less
Submitted 5 January, 2022; v1 submitted 28 January, 2021;
originally announced January 2021.
-
Detecting Adversarial Examples by Input Transformations, Defense Perturbations, and Voting
Authors:
Federico Nesti,
Alessandro Biondi,
Giorgio Buttazzo
Abstract:
Over the last few years, convolutional neural networks (CNNs) have proved to reach super-human performance in visual recognition tasks. However, CNNs can easily be fooled by adversarial examples, i.e., maliciously-crafted images that force the networks to predict an incorrect output while being extremely similar to those for which a correct output is predicted. Regular adversarial examples are not…
▽ More
Over the last few years, convolutional neural networks (CNNs) have proved to reach super-human performance in visual recognition tasks. However, CNNs can easily be fooled by adversarial examples, i.e., maliciously-crafted images that force the networks to predict an incorrect output while being extremely similar to those for which a correct output is predicted. Regular adversarial examples are not robust to input image transformations, which can then be used to detect whether an adversarial example is presented to the network. Nevertheless, it is still possible to generate adversarial examples that are robust to such transformations.
This paper extensively explores the detection of adversarial examples via image transformations and proposes a novel methodology, called \textit{defense perturbation}, to detect robust adversarial examples with the same input transformations the adversarial examples are robust to. Such a \textit{defense perturbation} is shown to be an effective counter-measure to robust adversarial examples.
Furthermore, multi-network adversarial examples are introduced. This kind of adversarial examples can be used to simultaneously fool multiple networks, which is critical in systems that use network redundancy, such as those based on architectures with majority voting over multiple CNNs. An extensive set of experiments based on state-of-the-art CNNs trained on the Imagenet dataset is finally reported.
△ Less
Submitted 16 August, 2021; v1 submitted 27 January, 2021;
originally announced January 2021.
-
A shape optimization problem on planar sets with prescribed topology
Authors:
L. Briani,
G. Buttazzo,
F. Prinari
Abstract:
We consider shape optimization problems involving functionals depending on perimeter, torsional rigidity and Lebesgue measure. The scaling free cost functionals are of the form $P(Ω)T^q(Ω)|Ω|^{-2q-1/2}$ and the class of admissible domains consists of two-dimensional open sets $Ω$ satisfying the topological constraints of having a prescribed number $k$ of bounded connected components of the complem…
▽ More
We consider shape optimization problems involving functionals depending on perimeter, torsional rigidity and Lebesgue measure. The scaling free cost functionals are of the form $P(Ω)T^q(Ω)|Ω|^{-2q-1/2}$ and the class of admissible domains consists of two-dimensional open sets $Ω$ satisfying the topological constraints of having a prescribed number $k$ of bounded connected components of the complementary set. A relaxed procedure is needed to have a well-posed problem and we show that when $q<1/2$ an optimal relaxed domain exists. When $q>1/2$ the problem is ill-posed and for $q=1/2$ the explicit value of the infimum is provided in the cases $k=0$ and $k=1$.
△ Less
Submitted 19 January, 2021; v1 submitted 8 January, 2021;
originally announced January 2021.
-
An application of the continuous Steiner symmetrization to Blaschke-Santaló diagrams
Authors:
Giuseppe Buttazzo,
Aldo Pratelli
Abstract:
In this paper we consider the so-called procedure of {\it Continuous Steiner Symmetrization}, introduced by Brock in \cite{bro95,bro00}. It transforms every domain $Ω\subset\subset\mathbb{R}^d$ into the ball kee** the volume fixed and letting the first eigenvalue and the torsion respectively decrease and increase. While this does not provide, in general, a $γ$-continuous map $t\mapstoΩ_t$, it ca…
▽ More
In this paper we consider the so-called procedure of {\it Continuous Steiner Symmetrization}, introduced by Brock in \cite{bro95,bro00}. It transforms every domain $Ω\subset\subset\mathbb{R}^d$ into the ball kee** the volume fixed and letting the first eigenvalue and the torsion respectively decrease and increase. While this does not provide, in general, a $γ$-continuous map $t\mapstoΩ_t$, it can be slightly modified so to obtain the $γ$-continuity for a $γ$-dense class of domains $Ω$, namely, the class of polyedral sets in $\mathbb{R}^d$. This allows to obtain a sharp characterization of the Blaschke-Santaló diagram of torsion and eigenvalue.
△ Less
Submitted 17 November, 2020;
originally announced November 2020.
-
Shape optimization problems for functionals with a boundary integral
Authors:
Giuseppe Buttazzo,
Francesco Paolo Maiale
Abstract:
We consider shape optimization problems for general integral functionals of the calculus of variations that may contain a boundary term. In particular, this class includes optimization problems governed by elliptic equations with a Robin condition on the free boundary. We show the existence of an optimal domain under rather general assumptions and we study the cases when the optimal domains are op…
▽ More
We consider shape optimization problems for general integral functionals of the calculus of variations that may contain a boundary term. In particular, this class includes optimization problems governed by elliptic equations with a Robin condition on the free boundary. We show the existence of an optimal domain under rather general assumptions and we study the cases when the optimal domains are open sets and have a finite perimeter.
△ Less
Submitted 22 July, 2020;
originally announced July 2020.
-
Optimal one-dimensional structures for the principal eigenvalue of two-dimensional domains
Authors:
Giuseppe Buttazzo,
Francesco Paolo Maiale
Abstract:
A shape optimization problem arising from the optimal reinforcement of a membrane by means of one-dimensional stiffeners or from the fastest cooling of a two-dimensional object by means of ``conducting wires'' is considered. The criterion we consider is the maximization of the first eigenvalue and the admissible classes of choices are the one of one-dimensional sets with prescribed total length, o…
▽ More
A shape optimization problem arising from the optimal reinforcement of a membrane by means of one-dimensional stiffeners or from the fastest cooling of a two-dimensional object by means of ``conducting wires'' is considered. The criterion we consider is the maximization of the first eigenvalue and the admissible classes of choices are the one of one-dimensional sets with prescribed total length, or the one where the constraint of being connected (or with an a priori bounded number of connected components) is added. The corresponding relaxed problems and the related existence results are described.
△ Less
Submitted 11 July, 2020;
originally announced July 2020.
-
Some inequalities involving perimeter and torsional rigidity
Authors:
L. Briani,
G. Buttazzo,
F. Prinari
Abstract:
We consider shape functionals of the form $F_q(Ω)=P(Ω)T^q(Ω)$ on the class of open sets of prescribed Lebesgue measure. Here $q>0$ is fixed, $P(Ω)$ denotes the perimeter of $Ω$ and $T(Ω)$ is the torsional rigidity of $Ω$. The minimization and maximization of $F_q(Ω)$ is considered on various classes of admissible domains $Ω$: in the class $\mathcal{A}_{all}$ of all domains, in the class…
▽ More
We consider shape functionals of the form $F_q(Ω)=P(Ω)T^q(Ω)$ on the class of open sets of prescribed Lebesgue measure. Here $q>0$ is fixed, $P(Ω)$ denotes the perimeter of $Ω$ and $T(Ω)$ is the torsional rigidity of $Ω$. The minimization and maximization of $F_q(Ω)$ is considered on various classes of admissible domains $Ω$: in the class $\mathcal{A}_{all}$ of all domains, in the class $\mathcal{A}_{convex}$ of convex domains, and in the class $\mathcal{A}_{thin}$ of thin domains.
△ Less
Submitted 6 July, 2020;
originally announced July 2020.
-
On capacity and torsional rigidity
Authors:
Michiel van den Berg,
Giuseppe Buttazzo
Abstract:
We investigate extremality properties of shape functionals which are products of Newtonian capacity $\cp(\overline{\Om})$, and powers of the torsional rigidity $T(\Om)$, for an open set $\Om\subset \R^d$ with compact closure $\overline{\Om}$, and prescribed Lebesgue measure. It is shown that if $\Om$ is convex then $\cp(\overline{\Om})T^q(\Om)$ is (i) bounded from above if and only if $q\ge 1$, an…
▽ More
We investigate extremality properties of shape functionals which are products of Newtonian capacity $\cp(\overline{\Om})$, and powers of the torsional rigidity $T(\Om)$, for an open set $\Om\subset \R^d$ with compact closure $\overline{\Om}$, and prescribed Lebesgue measure. It is shown that if $\Om$ is convex then $\cp(\overline{\Om})T^q(\Om)$ is (i) bounded from above if and only if $q\ge 1$, and (ii) bounded from below and away from $0$ if and only if $q\le \frac{d-2}{2(d-1)}$. Moreover a convex maximiser for the product exists if either $q>1$, or $d=3$ and $q=1$. A convex minimiser exists for $q< \frac{d-2}{2(d-1)}$. If $q\le 0$, then the product is minimised among all bounded sets by a ball of measure $1$.
△ Less
Submitted 4 September, 2020; v1 submitted 13 January, 2020;
originally announced January 2020.
-
On the relations between principal eigenvalue and torsional rigidity
Authors:
Michiel van den Berg,
Giuseppe Buttazzo,
Aldo Pratelli
Abstract:
We consider the problem of minimising or maximising the quantity $λ(Ø)T^q(Ø)$ on the class of open sets of prescribed Lebesgue measure. Here $q>0$ is fixed, $λ(Ø)$ denotes the first eigenvalue of the Dirichlet Laplacian on $H^1_0(Ø)$, while $T(Ø)$ is the torsional rigidity of $Ø$. The optimisation problem above is considered in the class of {\it all domains} $Ø$, in the class of {\it convex domain…
▽ More
We consider the problem of minimising or maximising the quantity $λ(Ø)T^q(Ø)$ on the class of open sets of prescribed Lebesgue measure. Here $q>0$ is fixed, $λ(Ø)$ denotes the first eigenvalue of the Dirichlet Laplacian on $H^1_0(Ø)$, while $T(Ø)$ is the torsional rigidity of $Ø$. The optimisation problem above is considered in the class of {\it all domains} $Ø$, in the class of {\it convex domains} $Ø$, and in the class of {\it thin domains}. The full Blaschke-Santaló diagram for $λ(Ø)$ and $T(Ø)$ is obtained in dimension one, while for higher dimensions we provide some bounds.
△ Less
Submitted 14 November, 2019; v1 submitted 31 October, 2019;
originally announced October 2019.
-
On existence of optimal potentials on unbounded domains
Authors:
Giuseppe Buttazzo,
Juan Casado-díaz,
Faustino Maestre
Abstract:
We consider elliptic equations of Schrödinger type with a right-hand side fixed and with the linear part of order zero given by a potential V . The main goal is to study the optimization problem for an integral cost depending on the solution uV , when V varies in a suitable class of admissible potentials. These problems can be seen as the natural extension of shape optimization problems to the fra…
▽ More
We consider elliptic equations of Schrödinger type with a right-hand side fixed and with the linear part of order zero given by a potential V . The main goal is to study the optimization problem for an integral cost depending on the solution uV , when V varies in a suitable class of admissible potentials. These problems can be seen as the natural extension of shape optimization problems to the framework of potentials. The main result is an existence theorem for optimal potentials, and the main difficulty is to work in the whole Euclidean space Rd, which implies a lack of compactness in several crucial points. In the last section we present some numerical simulations.
△ Less
Submitted 13 September, 2019;
originally announced September 2019.
-
Relaxed multi-marginal costs and quantization effects
Authors:
Guy Bouchitté,
Giuseppe Buttazzo,
Thierry Champion,
Luigi De Pascale
Abstract:
We propose a duality theory for multi-marginal repulsive cost that appear in optimal transport problems arising in Density Functional Theory. The related optimization problems involve probabilities on the entire space and, as minimizing sequences may lose mass at infinity, it is natural to expect relaxed solutions which are sub-probabilities. We first characterize the $N$-marginals relaxed cost in…
▽ More
We propose a duality theory for multi-marginal repulsive cost that appear in optimal transport problems arising in Density Functional Theory. The related optimization problems involve probabilities on the entire space and, as minimizing sequences may lose mass at infinity, it is natural to expect relaxed solutions which are sub-probabilities. We first characterize the $N$-marginals relaxed cost in terms of a stratification formula which takes into account all $k$ interactions with $k\le N$. We then develop a duality framework involving continuous functions vanishing at infinity and deduce primal-dual necessary and sufficient optimality conditions Next we prove the existence and the regularity of an optimal dual potential under very mild assumptions. In the last part of the paper, we apply our results to a minimization problem involving a given continuous potential and we give evidence of a mass quantization effect for optimal solutions.
△ Less
Submitted 19 July, 2019;
originally announced July 2019.
-
Optimal reinforcing networks for elastic membranes
Authors:
Giovanni Alberti,
Giuseppe Buttazzo,
Serena Guarino Lo Bianco,
Edouard Oudet
Abstract:
In this paper we study the optimal reinforcement of an elastic membrane, fixed at its boundary, by means of a network (connected onedimensional structure), that has to be found in a suitable admissible class. We show the existence of an optimal solution that may present multiplicities, that is regions where the optimal structure overlaps. Some numerical simulations are shown to confirm this issue…
▽ More
In this paper we study the optimal reinforcement of an elastic membrane, fixed at its boundary, by means of a network (connected onedimensional structure), that has to be found in a suitable admissible class. We show the existence of an optimal solution that may present multiplicities, that is regions where the optimal structure overlaps. Some numerical simulations are shown to confirm this issue and to illustrate the complexity of the optimal network when their total length becomes large.
△ Less
Submitted 21 May, 2019; v1 submitted 3 April, 2019;
originally announced April 2019.
-
Dissociating limit in Density Functional Theory with Coulomb optimal transport cost
Authors:
Guy Bouchitté,
Giuseppe Buttazzo,
Thierry Champion,
Luigi De Pascale
Abstract:
In the framework of Density Functional Theory with Strongly Correlated Electrons we consider the so called bond dissociating limit for the energy of an aggregate of atoms. We show that the multi-marginals optimal transport cost with Coulombian electron-electron repulsion may correctly describe the dissociation effect. The variational limit is completely calculated in the case of N=2 electrons. The…
▽ More
In the framework of Density Functional Theory with Strongly Correlated Electrons we consider the so called bond dissociating limit for the energy of an aggregate of atoms. We show that the multi-marginals optimal transport cost with Coulombian electron-electron repulsion may correctly describe the dissociation effect. The variational limit is completely calculated in the case of N=2 electrons. The theme of fractional number of electrons appears naturally and brings into play the question of optimal partial transport cost. A plan is outlined to complete the analysis which involves the study of the relaxation of optimal transport cost with respect to the weak* convergence of measures.
△ Less
Submitted 29 November, 2018;
originally announced November 2018.
-
Optimal shapes for general integral functionals
Authors:
Giuseppe Buttazzo,
Harish Shrivastava
Abstract:
We consider shape optimization problems for general integral functionals of the calculus of variations, defined on a domain $Ω$ that varies over all subdomains of a given bounded domain $D$ of ${\bf R}^d$. We show in a rather elementary way the existence of a solution that is in general a quasi open set. Under very mild conditions we show that the optimal domain is actually open and with finite pe…
▽ More
We consider shape optimization problems for general integral functionals of the calculus of variations, defined on a domain $Ω$ that varies over all subdomains of a given bounded domain $D$ of ${\bf R}^d$. We show in a rather elementary way the existence of a solution that is in general a quasi open set. Under very mild conditions we show that the optimal domain is actually open and with finite perimeter. Some counterexamples show that in general this does not occur.
△ Less
Submitted 25 March, 2018;
originally announced March 2018.
-
Optimal potentials for problems with changing sing data
Authors:
Giuseppe Buttazzo,
Faustino Maestre,
Bozhidar Velichkov
Abstract:
We consider optimal control problems where the state equation is an elliptic PDE of a Schrödinger type, governed by the Laplace operator $-Δ$ with the addition of a potential V, and the control is the potential V itself, that may vary in a suitable admissible class. In a previous paper (Ref. [7]) an existence result was established under a monotonicity assumption on the cost functional, which occu…
▽ More
We consider optimal control problems where the state equation is an elliptic PDE of a Schrödinger type, governed by the Laplace operator $-Δ$ with the addition of a potential V, and the control is the potential V itself, that may vary in a suitable admissible class. In a previous paper (Ref. [7]) an existence result was established under a monotonicity assumption on the cost functional, which occurs if the data do not change sign. In the present paper this sign assumption is removed and the existence of an optimal potential is still valid. Several numerical simulations, made by FreeFem++, are shown
△ Less
Submitted 23 October, 2017;
originally announced October 2017.
-
Two optimization problems in thermal insulation
Authors:
Dorin Bucur,
Giuseppe Buttazzo,
Carlo Nitsch
Abstract:
We consider two optimization problems in thermal insulation: in both cases the goal is to find a thin layer around the boundary of the thermal body which gives the best insulation. The total mass of the insulating material is prescribed.. The first problem deals with the case in which a given heat source is present, while in the second one there are no heat sources and the goal is to have the slow…
▽ More
We consider two optimization problems in thermal insulation: in both cases the goal is to find a thin layer around the boundary of the thermal body which gives the best insulation. The total mass of the insulating material is prescribed.. The first problem deals with the case in which a given heat source is present, while in the second one there are no heat sources and the goal is to have the slowest decay of the temperature. In both cases an optimal distribution of the insulator around the thermal body exists; when the body has a circular symmetry, in the first case a constant heat source gives a constant thickness as the optimal solution, while surprisingly this is not the case in the second problem, where the circular symmetry of the optimal insulating layer depends on the total quantity of insulator at our disposal. A symmetry breaking occurs when this total quantity is below a certain threshold. Some numerical computations are also provided, together with a list of open questions.
△ Less
Submitted 28 August, 2017;
originally announced August 2017.
-
Numerical solution of a nonlinear eigenvalue problem arising in optimal insulation
Authors:
Sören Bartels,
Giuseppe Buttazzo
Abstract:
The optimal insulation of a heat conducting body by a thin film of variable thickness can be formulated as a nondifferentiable, nonlocal eigenvalue problem. The discretization and iterative solution for the reliable computation of corresponding eigenfunctions that determine the optimal layer thickness are addressed. Corresponding numerical experiments confirm the theoretical observation that a sym…
▽ More
The optimal insulation of a heat conducting body by a thin film of variable thickness can be formulated as a nondifferentiable, nonlocal eigenvalue problem. The discretization and iterative solution for the reliable computation of corresponding eigenfunctions that determine the optimal layer thickness are addressed. Corresponding numerical experiments confirm the theoretical observation that a symmetry breaking occurs for the case of small available insulation masses and provide insight in the geometry of optimal films. An experimental shape optimization indicates that convex bodies with one axis of symmetry have favorable insulation properties.
△ Less
Submitted 12 August, 2017;
originally announced August 2017.
-
A shape optimal control problem and its probabilistic counterpart
Authors:
Giuseppe Buttazzo,
Bozhidar Velichkov
Abstract:
In this paper we consider a shape optimization problem in which the data in the cost functional and in the state equation may change sign, and so no monotonicity assumption is satisfied. Nevertheless, we are able to prove that an optimal domain exists. We also deduce some necessary conditions of optimality for the optimal domain. The results are applied to show the existence of an optimal domain i…
▽ More
In this paper we consider a shape optimization problem in which the data in the cost functional and in the state equation may change sign, and so no monotonicity assumption is satisfied. Nevertheless, we are able to prove that an optimal domain exists. We also deduce some necessary conditions of optimality for the optimal domain. The results are applied to show the existence of an optimal domain in the case where the cost functional is completely identified, while the right-hand side in the state equation is only known up to a probability P in the space $L^2$.
△ Less
Submitted 21 April, 2017;
originally announced April 2017.
-
Sharp estimates for the anisotropic torsional rigidity and the principal frequency
Authors:
Giuseppe Buttazzo,
Serena Guarino Lo Bianco,
Michele Marini
Abstract:
In this paper we generalize some classical estimates involving the torsional rigidity and the principal frequency of a convex domain to a class of functionals related to some anisotropic non linear operators.
In this paper we generalize some classical estimates involving the torsional rigidity and the principal frequency of a convex domain to a class of functionals related to some anisotropic non linear operators.
△ Less
Submitted 27 March, 2017;
originally announced March 2017.
-
Optimal Control Problems in Transport Dynamics
Authors:
Mattia Bongini,
Giuseppe Buttazzo
Abstract:
In the present paper we deal with an optimal control problem related to a model in population dynamics; more precisely, the goal is to modify the behavior of a given density of individuals via another population of agents interacting with the first. The cost functional to be minimized to determine the dynamics of the second population takes into account the desired target or configuration to be re…
▽ More
In the present paper we deal with an optimal control problem related to a model in population dynamics; more precisely, the goal is to modify the behavior of a given density of individuals via another population of agents interacting with the first. The cost functional to be minimized to determine the dynamics of the second population takes into account the desired target or configuration to be reached as well as the quantity of control agents. Several applications may fall into this framework, as for instance driving a mass of pedestrian in (or out of) a certain location; influencing the stock market by acting on a small quantity of key investors; controlling a swarm of unmanned aerial vehicles by means of few piloted drones.
△ Less
Submitted 23 September, 2016;
originally announced September 2016.
-
Continuity and estimates for multimarginal optimal transportation problems with singular costs
Authors:
Giuseppe Buttazzo,
Thierry Champion,
Luigi De Pascale
Abstract:
We consider some repulsive multimarginal optimal transportation problems which include, as a particular case, the Coulomb cost. We prove a regularity property of the minimizers (optimal transportation plan) from which we deduce existence and some basic regularity of a maximizer for the dual problem (Kantorovich potential). This is then applied to obtain some estimates of the cost and to the study…
▽ More
We consider some repulsive multimarginal optimal transportation problems which include, as a particular case, the Coulomb cost. We prove a regularity property of the minimizers (optimal transportation plan) from which we deduce existence and some basic regularity of a maximizer for the dual problem (Kantorovich potential). This is then applied to obtain some estimates of the cost and to the study of continuity properties.
△ Less
Submitted 31 August, 2016;
originally announced August 2016.
-
Worst-case shape optimization for the Dirichlet energy
Authors:
José Carlos Bellido,
Giuseppe Buttazzo,
Bozhidar Velichkov
Abstract:
We consider the optimization problem for a shape cost functional $F(Ω,f)$ which depends on a domain $Ω$ varying in a suitable admissible class and on a "right-hand side" $f$. More precisely, the cost functional $F$ is given by an integral which involves the solution $u$ of an elliptic PDE in $Ω$ with right-hand side $f$; the boundary conditions considered are of the Dirichlet type. When the functi…
▽ More
We consider the optimization problem for a shape cost functional $F(Ω,f)$ which depends on a domain $Ω$ varying in a suitable admissible class and on a "right-hand side" $f$. More precisely, the cost functional $F$ is given by an integral which involves the solution $u$ of an elliptic PDE in $Ω$ with right-hand side $f$; the boundary conditions considered are of the Dirichlet type. When the function $f$ is only known up to some degree of uncertainty, our goal is to obtain the existence of an optimal shape in the worst possible situation. Some numerical simulations are provided, showing the difference in the optimal shape between the case when $f$ is perfectly known and the case when only the worst situation is optimized.
△ Less
Submitted 17 May, 2016;
originally announced May 2016.
-
Thin elastic plates supported over small areas. II. Variational-asymptotic models
Authors:
G. Buttazzo,
G. Cardone,
S. A. Nazarov
Abstract:
An asymptotic analysis is performed for thin anisotropic elastic plate clamped along its lateral side and also supported at a small area $θ_{h}$ of one base with diameter of the same order as the plate thickness $h\ll1.$ A three-dimensional boundary layer in the vicinity of the support $θ_{h}$ is involved into the asymptotic form which is justified by means of the previously derived weighted inequ…
▽ More
An asymptotic analysis is performed for thin anisotropic elastic plate clamped along its lateral side and also supported at a small area $θ_{h}$ of one base with diameter of the same order as the plate thickness $h\ll1.$ A three-dimensional boundary layer in the vicinity of the support $θ_{h}$ is involved into the asymptotic form which is justified by means of the previously derived weighted inequality of Korn's type provides an error estimate with the bound $ ch^{1/2} \left | \ln h\right| .$ Ignoring this boundary layer effect reduces the precision order down to $\left| \ln h\right| ^{-1/2}.$ A two-dimensional variational-asymptotic model of the plate is proposed within the theory of self-adjoint extensions of differential operators. The only characteristics of the boundary layer, namely the elastic logarithmic potential matrix of size $4\times4,$ is involved into the model which however keeps the precision order $h^{1/2}\left| \ln h\right| $ in certain norms. Several formulations and applications of the model are discussed.
△ Less
Submitted 19 January, 2016;
originally announced January 2016.
-
Symmetry breaking for a problem in optimal insulation
Authors:
Dorin Bucur,
Giuseppe Buttazzo,
Carlo Nitsch
Abstract:
We consider the problem of optimally insulating a given domain $Ω$ of ${\mathbb{R}}^d$; this amounts to solve a nonlinear variational problem, where the optimal thickness of the insulator is obtained as the boundary trace of the solution. We deal with two different criteria of optimization: the first one consists in the minimization of the total energy of the system, while the second one involves…
▽ More
We consider the problem of optimally insulating a given domain $Ω$ of ${\mathbb{R}}^d$; this amounts to solve a nonlinear variational problem, where the optimal thickness of the insulator is obtained as the boundary trace of the solution. We deal with two different criteria of optimization: the first one consists in the minimization of the total energy of the system, while the second one involves the first eigenvalue of the related differential operator. Surprisingly, the second optimization problem presents a symmetry breaking in the sense that for a ball the optimal thickness is nonsymmetric when the total amount of insulator is small enough. In the last section we discuss the shape optimization problem which is obtained letting $Ω$ to vary too.
△ Less
Submitted 9 January, 2016;
originally announced January 2016.
-
Thin elastic plates supported over small areas. I. Korn's inequalities and boundary layers
Authors:
G. Buttazzo,
G. Cardone,
S. A. Nazarov
Abstract:
A thin anisotropic elastic plate clamped along its lateral side and also supported at a small area $θ_{h}$ of one base is considered; the diameter of $θ_{h}$ is of the same order as the plate relative thickness $h\ll1$. In addition to the standard Kirchhoff model with the Sobolev point condition, a three-dimensional boundary layer is investigated in the vicinity of the support $θ_{h}$, which with…
▽ More
A thin anisotropic elastic plate clamped along its lateral side and also supported at a small area $θ_{h}$ of one base is considered; the diameter of $θ_{h}$ is of the same order as the plate relative thickness $h\ll1$. In addition to the standard Kirchhoff model with the Sobolev point condition, a three-dimensional boundary layer is investigated in the vicinity of the support $θ_{h}$, which with the help of the derived weighted inequality of Korn's type, will provide an error estimate with the bound $ch^{1/2}|\ln h|$. Ignoring this boundary layer effect reduces the precision order down to $|\ln h|^{-1/2}$.
△ Less
Submitted 14 July, 2015;
originally announced July 2015.
-
A free boundary problem arising in PDE optimization
Authors:
Giuseppe Buttazzo,
Edouard Oudet,
Bozhidar Velichkov
Abstract:
A free boundary problem arising from the optimal reinforcement of a membrane or from the reduction of traffic congestion is considered; it is of the form $$\sup_{\int_Dθ\,dx=m}\ \inf_{u\in H^1_0(D)}\int_D\Big(\frac{1+θ}{2}|\nabla u|^2-fu\Big)\,dx.$$ We prove the existence of an optimal reinforcement $θ$ and that it has some higher integrability properties. We also provide some numerical computatio…
▽ More
A free boundary problem arising from the optimal reinforcement of a membrane or from the reduction of traffic congestion is considered; it is of the form $$\sup_{\int_Dθ\,dx=m}\ \inf_{u\in H^1_0(D)}\int_D\Big(\frac{1+θ}{2}|\nabla u|^2-fu\Big)\,dx.$$ We prove the existence of an optimal reinforcement $θ$ and that it has some higher integrability properties. We also provide some numerical computations for $θ$ and $u$.
△ Less
Submitted 30 May, 2015;
originally announced June 2015.
-
Optimal design problems for Schrödinger operators with noncompact resolvents
Authors:
Guy Bouchitté,
Giuseppe Buttazzo
Abstract:
We consider optimization problems for cost functionals which depend on the negative spectrum of Schrödinger operators of the form $-Δ+V(x)$, where $V$ is a potential, with prescribed compact support, which has to be determined. Under suitable assumptions the existence of an optimal potential is shown. This can be applied to interesting cases such as costs functions involving finitely many negative…
▽ More
We consider optimization problems for cost functionals which depend on the negative spectrum of Schrödinger operators of the form $-Δ+V(x)$, where $V$ is a potential, with prescribed compact support, which has to be determined. Under suitable assumptions the existence of an optimal potential is shown. This can be applied to interesting cases such as costs functions involving finitely many negative eigenvalues.
△ Less
Submitted 31 January, 2015;
originally announced February 2015.
-
The spectral drop problem
Authors:
Giuseppe Buttazzo,
Bozhidar Velichkov
Abstract:
We consider spectral optimization problems of the form $$\min\Big\{λ_1(Ω;D):\ Ω\subset D,\ |Ω|=1\Big\},$$ where $D$ is a given subset of the Euclidean space $\mathbb{R}^d$. Here $λ_1(Ω;D)$ is the first eigenvalue of the Laplace operator $-Δ$ with Dirichlet conditions on $\partialΩ\cap D$ and Neumann or Robin conditions on $\partialΩ\cap\partial D$. The equivalent variational formulation…
▽ More
We consider spectral optimization problems of the form $$\min\Big\{λ_1(Ω;D):\ Ω\subset D,\ |Ω|=1\Big\},$$ where $D$ is a given subset of the Euclidean space $\mathbb{R}^d$. Here $λ_1(Ω;D)$ is the first eigenvalue of the Laplace operator $-Δ$ with Dirichlet conditions on $\partialΩ\cap D$ and Neumann or Robin conditions on $\partialΩ\cap\partial D$. The equivalent variational formulation $$λ_1(Ω;D)=\min\left\{\int_Ω|\nabla u|^2\,dx+k\int_{\partial D}u^2\,d\mathcal{H}^{d-1}\ :\ u\in H^1(D),\ u=0\hbox{ on }\partialΩ\cap D,\ \|u\|_{L^2(Ω)}=1\right\}$$ reminds the classical drop problems, where the first eigenvalue replaces the total variation functional. We prove an existence result for general shape cost functionals and we show some qualitative properties of the optimal domains.
△ Less
Submitted 6 June, 2014;
originally announced June 2014.
-
Optimal regions for congested transport
Authors:
Giuseppe Buttazzo,
Guillaume Carlier,
Serena Guarino Lo Bianco
Abstract:
We consider a given region $Ω$ where the traffic flows according to two regimes: in a region $C$ we have a low congestion, where in the remaining part $Ω\setminus C$ the congestion is higher. The two congestion functions $H_1$ and $H_2$ are given, but the region $C$ has to be determined in an optimal way in order to minimize the total transportation cost. Various penalization terms on $C$ are cons…
▽ More
We consider a given region $Ω$ where the traffic flows according to two regimes: in a region $C$ we have a low congestion, where in the remaining part $Ω\setminus C$ the congestion is higher. The two congestion functions $H_1$ and $H_2$ are given, but the region $C$ has to be determined in an optimal way in order to minimize the total transportation cost. Various penalization terms on $C$ are considered and some numerical computations are shown.
△ Less
Submitted 3 April, 2014;
originally announced April 2014.
-
Optimization problems involving the first Dirichlet eigenvalue and the torsional rigidity
Authors:
Giuseppe Buttazzo,
Michiel van den Berg,
Bozhidar Velichkov
Abstract:
We present some open problems and obtain some partial results for spectral optimization problems involving measure, torsional rigidity and first Dirichlet eigenvalue.
We present some open problems and obtain some partial results for spectral optimization problems involving measure, torsional rigidity and first Dirichlet eigenvalue.
△ Less
Submitted 31 October, 2014; v1 submitted 1 March, 2014;
originally announced March 2014.
-
Improved energy bounds for Schrödinger operators
Authors:
Lorenzo Brasco,
Giuseppe Buttazzo
Abstract:
Given a potential $V$ and the associated Schrödinger operator $-Δ+V$, we consider the problem of providing sharp upper and lower bound on the energy of the operator. It is known that if for example $V$ or $V^{-1}$ enjoys suitable summability properties, the problem has a positive answer. In this paper we show that the corresponding isoperimetric-like inequalities can be improved by means of quanti…
▽ More
Given a potential $V$ and the associated Schrödinger operator $-Δ+V$, we consider the problem of providing sharp upper and lower bound on the energy of the operator. It is known that if for example $V$ or $V^{-1}$ enjoys suitable summability properties, the problem has a positive answer. In this paper we show that the corresponding isoperimetric-like inequalities can be improved by means of quantitative stability estimates.
△ Less
Submitted 15 July, 2014; v1 submitted 24 February, 2014;
originally announced February 2014.
