Les surfaces de subdivision fournissent une méthode alternative prometteuse dans la modélisation géométrique, et ont des avantages sur la représentation classique de trimmed-NURBS, en particulier dans la modélisation de surfaces lisses par morceaux. Dans ce mémoire, nous considérons le problème des opérations géométriques sur les surfaces de subdivision, avec l'exigence stricte de forme topologique correcte. Puisque ce problème peut être mal conditionné, nous proposons une approche pour la gestion de l'incertitude qui existe dans le calcul géométrique. Nous exigeons l'exactitude des informations topologiques lorsque l'on considère la nature de robustesse du problème des opérations géométriques sur les modèles de solides, et il devient clair que le problème peut être mal conditionné en présence de l'incertitude qui est omniprésente dans les données. Nous proposons donc une approche interactive de gestion de l'incertitude des opérations géométriques, dans le cadre d'un calcul basé sur la norme IEEE arithmétique et la modélisation en surfaces de subdivision. Un algorithme pour le problème planar-cut est alors présenté qui a comme but de satisfaire à l'exigence topologique mentionnée ci-dessus. / Subdivision surfaces are a promising alternative method for geometric modelling, and have some important advantages over the classical representation of trimmed-NURBS, especially in modelling piecewise smooth surfaces. In this thesis, we consider the problem of geometric operations on subdivision surfaces with the strict requirement of correct topological form, and since this problem may be ill-conditioned, we propose an approach for managing uncertainty that exists inherently in geometric computation. We take into account the requirement of the correctness of topological information when considering the nature of robustness for the problem of geometric operations on solid models, and it becomes clear that the problem may be ill-conditioned in the presence of uncertainty that is ubiquitous in the data. Starting from this point, we propose an interactive approach of managing uncertainty of geometric operations, in the context of computation using the standard IEEE arithmetic and modelling using a subdivision-surface representation. An algorithm for the planar-cut problem is then presented, which has as its goal the satisfaction of the topological requirement mentioned above.

Cette thèse est dédiée au développement d’algorithmes génériques pour l’observation ensembliste de l’état continu et du mode discret des systèmes dynamiques hybrides dans le but de réaliser la détection de défauts. Cette thèse est organisée en deux grandes parties. Dans la première partie, nous avons proposé une méthode rapide et efficace pour le passage ensembliste des gardes. Elle consiste à procéder à la bissection dans la seule direction du temps et ensuite faire collaborer plusieurs contracteurs simultanément pour réduire le domaine des vecteurs d’état localisés sur la garde, durant la tranche de temps étudiée. Ensuite, nous avons proposé une méthode pour la fusion des trajectoires basée sur l'utilisation des zonotopes. Ces méthodes, utilisées conjointement, nous ont permis de caractériser de manière garantie l'ensemble des trajectoires d'état hybride engendrées par un système dynamique hybride incertain sur un horizon de temps fini. La deuxième partie de la thèse aborde les méthodes ensemblistes pour l'estimation de paramètres et pour l'estimation d'état hybride (mode et état continu) dans un contexte à erreurs bornées. Nous avons commencé en premier lieu par décrire les méthodes de détection de défauts dans les systèmes hybrides en utilisant une approche paramétrique et une approche observateur hybride. Ensuite, nous avons décrit deux méthodes permettant d’effectuer les tâches de détection de défauts. Nous avons proposé une méthode basée sur notre méthode d'atteignabilité hybride non linéaire et un algorithme de partitionnement que nous avons nommé SIVIA-H pour calculer de manière garantie l'ensemble des paramètres compatibles avec le modèle hybride, les mesures et avec les bornes d’erreurs. Ensuite, pour l'estimation d'état hybride, nous avons proposé une méthode basée sur un prédicteurcorrecteur construit au dessus de notre méthode d'atteignabilité hybride non linéaire. / This thesis is dedicated to the development of generic algorithms for the set-membership observation of the continuous state and the discrete mode of hybrid dynamical systems in order to achieve fault detection. This thesis is organized into two parts. In the first part, we have proposed a fast and effective method for the set-membership guard crossing. It consists in carrying out bisection in the time direction only and then makes several contractors working simultaneously to reduce the domain of state vectors located on the guard during the study time slot. Then, we proposed a method for merging trajectories based on zonotopic enclosures. These methods, used together, allowed us to characterize in a guaranteed way the set of all hybrid state trajectories generated by an uncertain hybrid dynamical system on a finite time horizon. The second part focuses on set-membership methods for the parameters or the hybrid state (mode and continuous state) of a hybrid dynamical system in a bounded error framework. We started first by describing fault detection methods for hybrid systems using the parametric approach and the hybrid observer approach. Then, we have described two methods for performing fault detection tasks. We have proposed a method for computing in a guaranteed way all the parameters consistent with the hybrid dynamical model, the actual data and the prior error bound, by using our nonlinear hybrid reachability method and an algorithm for partition which we denote SIVIA-H. Then, for hybrid state estimation, we have proposed a method based on a predictor-corrector, which is also built on top of our non-linear method for hybrid reachability.

Fundamentada a importância da utilização da Teoria dos Intervalos em computação científica, é realizada uma revisão da Teoria Clássica dos Intervalos, com críticas sobre as incompatibilidades encontradas como motivos de diversas dificuldades para desenvolvimento da própria teoria e, consequentemente, das Técnicas Intervalares. É desenvolvida uma nova abordagem para a Teoria dos Intervalos de acordo com a Teoria dos Domínios e a proposta de [ACI 89], obtendo-se os Domínios Intervalares da Matemática Computacional. Introduz-se uma topologia (Topologia de Scott) compatível com a idéia de aproximação, gerando uma ordem de informação, isto é, para quaisquer intervalos x e y, diz-se que se x -c y , então y fornece mais (no mínimo tanto quanto) informação, sobre um real r, do que x. Prova-se que esta ordem de informação induz uma topologia To (topologia de Scott) , que é mais adequada para uma teoria computacional que a topologia da Hausdorff introduzida por Moore [MOO 66]. Cada número real r é aproximado por intervalos de extremos racionais, os intervalos de informação, que constituem o espaço de informação II(Q), superando assim a regressão infinita da abordagem clássica. Pode-se dizer que todo real r é o supremo de uma cadeia de intervalos com extremos racionais “encaixados”. Assim, os reais são os elementos totais de um domínio contínuo, chamado de Domínio dos Intervalos Reais Parciais, cuja base é o espaço de informação II (Q). Cada função contínua da Análise Real é o limite de sequências de funções contínuas entre elementos da base do domínio. Toda função contínua nestes domínios constitui uma função monotônica na base e é completamente representada em termos finitos. É introduzida uma quasi-métrica que induz uma topologia compatível com esta abordagem e provê as propriedades quantitativas, além de possibilitar a utilização da noção de sequências, limites etc, sem que se precise recorrer a conceitos mais complexos. Desenvolvem-se uma aritmética, critérios de aproximação e os conceito de intervalo ponto médio, intervalo valor absoluto e intervalo diâmetro, conceitos compatíveis com esta abordagem. São acrescentadas as operações de união, interseção e as unárias. Apresenta-se um amplo estudo sobre a função intervalar e a inclusão de imagens de funções, com ênfase na obtenção de uma extensão intervalar natural contínua. Esta é uma abordagem de lógica construtiva e computacional. / The importance of Interval Theory for scientific computation is emphasized. A review of the Classical Theory is macle, including a discussion about some incompatibities that cause problems in developing interval algorithms. A new approach to the Interval Theory is developed in the light of the Theory of Domains and according to the ideas by Acióly [ACI 89], getting the Interval Domains of Computational Mathematics. It is introduced a topology (Scott Topology), which is associated with the idea of approximation, generating an information order, that is, for any intervals x and y one says that if x -c y, then "the information given by y is better or at least equal than the one given by x". One proves that this information order induces a To topology (Scott's topology) which is more suitable for a computation theory than that of Hausdorff introduced by Moore [MOO 66]. This approach has the advantage of being both of constructive logic and computational. Each real number is approximated by intervals with rational bounds, named information intervals of the Information Space II(Q), eliminating the infinite regression found in the classical approach. One can say that every real a is the supreme of a chain of rational intervals. Then, the real numbers are the total elements of a continuous domain, named the Domain of the Partial Real Intervals, whose basis is the information space II (Q). Each continuous function in the Real Analysis is the limit of sequences of continuous functions among any elements which belong to the base of the domain. In these same domains, each continuous function is monotonic on the base and it is completely represented by finite terms. It is introduced a quasi-metric that leads to a compatible topology and supplies the quantitative properties. An arithmetic, some approximation criteria, the concepts of mean point interval, absolute value interval and width interval are developed and set operations are added. The ideas of interval functions and the inclusion of ranges of functions are also presented, and a continuous natural interval extension is obtained.

Fundamentada a importância da utilização da Teoria dos Intervalos em computação científica, é realizada uma revisão da Teoria Clássica dos Intervalos, com críticas sobre as incompatibilidades encontradas como motivos de diversas dificuldades para desenvolvimento da própria teoria e, consequentemente, das Técnicas Intervalares. É desenvolvida uma nova abordagem para a Teoria dos Intervalos de acordo com a Teoria dos Domínios e a proposta de [ACI 89], obtendo-se os Domínios Intervalares da Matemática Computacional. Introduz-se uma topologia (Topologia de Scott) compatível com a idéia de aproximação, gerando uma ordem de informação, isto é, para quaisquer intervalos x e y, diz-se que se x -c y , então y fornece mais (no mínimo tanto quanto) informação, sobre um real r, do que x. Prova-se que esta ordem de informação induz uma topologia To (topologia de Scott) , que é mais adequada para uma teoria computacional que a topologia da Hausdorff introduzida por Moore [MOO 66]. Cada número real r é aproximado por intervalos de extremos racionais, os intervalos de informação, que constituem o espaço de informação II(Q), superando assim a regressão infinita da abordagem clássica. Pode-se dizer que todo real r é o supremo de uma cadeia de intervalos com extremos racionais “encaixados”. Assim, os reais são os elementos totais de um domínio contínuo, chamado de Domínio dos Intervalos Reais Parciais, cuja base é o espaço de informação II (Q). Cada função contínua da Análise Real é o limite de sequências de funções contínuas entre elementos da base do domínio. Toda função contínua nestes domínios constitui uma função monotônica na base e é completamente representada em termos finitos. É introduzida uma quasi-métrica que induz uma topologia compatível com esta abordagem e provê as propriedades quantitativas, além de possibilitar a utilização da noção de sequências, limites etc, sem que se precise recorrer a conceitos mais complexos. Desenvolvem-se uma aritmética, critérios de aproximação e os conceito de intervalo ponto médio, intervalo valor absoluto e intervalo diâmetro, conceitos compatíveis com esta abordagem. São acrescentadas as operações de união, interseção e as unárias. Apresenta-se um amplo estudo sobre a função intervalar e a inclusão de imagens de funções, com ênfase na obtenção de uma extensão intervalar natural contínua. Esta é uma abordagem de lógica construtiva e computacional. / The importance of Interval Theory for scientific computation is emphasized. A review of the Classical Theory is macle, including a discussion about some incompatibities that cause problems in developing interval algorithms. A new approach to the Interval Theory is developed in the light of the Theory of Domains and according to the ideas by Acióly [ACI 89], getting the Interval Domains of Computational Mathematics. It is introduced a topology (Scott Topology), which is associated with the idea of approximation, generating an information order, that is, for any intervals x and y one says that if x -c y, then "the information given by y is better or at least equal than the one given by x". One proves that this information order induces a To topology (Scott's topology) which is more suitable for a computation theory than that of Hausdorff introduced by Moore [MOO 66]. This approach has the advantage of being both of constructive logic and computational. Each real number is approximated by intervals with rational bounds, named information intervals of the Information Space II(Q), eliminating the infinite regression found in the classical approach. One can say that every real a is the supreme of a chain of rational intervals. Then, the real numbers are the total elements of a continuous domain, named the Domain of the Partial Real Intervals, whose basis is the information space II (Q). Each continuous function in the Real Analysis is the limit of sequences of continuous functions among any elements which belong to the base of the domain. In these same domains, each continuous function is monotonic on the base and it is completely represented by finite terms. It is introduced a quasi-metric that leads to a compatible topology and supplies the quantitative properties. An arithmetic, some approximation criteria, the concepts of mean point interval, absolute value interval and width interval are developed and set operations are added. The ideas of interval functions and the inclusion of ranges of functions are also presented, and a continuous natural interval extension is obtained.

L'arithmétique par intervalles permet de calculer et simultanément vérifier des résultats. Cependant, une application naïve de cette arithmétique conduit à un encadrement grossier des résultats. De plus, de tels calculs peuvent être lents.Nous proposons des algorithmes précis et des implémentations efficaces, utilisant l'arithmétique par intervalles, dans le domaine de l'algèbre linéaire. Deux problèmes sont abordés : la multiplication de matrices à coefficients intervalles et la résolution vérifiée de systèmes linéaires. Pour le premier problème, nous proposons deux algorithmes qui offrent de bons compromis entre vitesse et précision. Pour le second problème, nos principales contributions sont d'une part une technique de relaxation, qui réduit substantiellement le temps d'exécution de l'algorithme, et d'autre part l'utilisation d'une précision étendue en quelques portions bien choisies de l'algorithme, afin d'obtenir rapidement une grande précision. / Interval arithmetic is a means to compute verified results. However, a naive use of interval arithmetic does not provide accurate enclosures of the exact results. Moreover, interval arithmetic computations can be time-consuming. We propose several accurate algorithms and efficient implementations in verified linear algebra using interval arithmetic. Two fundamental problems are addressed, namely the multiplication of interval matrices and the verification of a floating-point solution of a linear system. For the first problem, we propose two algorithms which offer new tradeoffs between speed and accuracy. For the second problem, which is the verification of the solution of a linear system, our main contributions are twofold. First, we introduce a relaxation technique, which reduces drastically the execution time of the algorithm. Second, we propose to use extended precision for few, well-chosen parts of the computations, to gain accuracy without losing much in term of execution time.

L'objet est de comparer des algorithmes de multiplication de matrices à coefficients intervalles et leurs implémentations.Le premier axe est la mesure de la précision numérique. Les précédentes analyses d'erreur se limitent à établir une borne sur la surestimation du rayon du résultat en négligeant les erreurs dues au calcul en virgule flottante. Après examen des différentes possibilités pour quantifier l'erreur d'approximation entre deux intervalles, l'erreur d'arrondi est intégrée dans l'erreur globale. À partir de jeux de données aléatoires, la dispersion expérimentale de l'erreur globale permet d'éclairer l'importance des différentes erreurs (de méthode et d'arrondi) en fonction de plusieurs facteurs : valeur et homogénéité des précisions relatives des entrées, dimensions des matrices, précision de travail. Cette démarche conduit à un nouvel algorithme moins coûteux et tout aussi précis dans certains cas déterminés.Le deuxième axe est d'exploiter le parallélisme des opérations. Les implémentations précédentes se ramènent à des produits de matrices de nombres flottants. Pour contourner les limitations d'une telle approche sur la validité du résultat et sur la capacité à monter en charge, je propose une implémentation par blocs réalisée avec des threads OpenMP qui exécutent des noyaux de calcul utilisant les instructions vectorielles. L'analyse des temps d'exécution sur une machine de 4 octo-coeurs montre que les coûts de calcul sont du même ordre de grandeur sur des matrices intervalles et numériques de même dimension et que l'implémentation par bloc passe mieux à l'échelle que l'implémentation avec plusieurs appels aux routines BLAS. / This work aims at determining suitable scopes for several algorithms of interval matrices multiplication.First, we quantify the numerical quality. Former error analyses of interval matrix products establish bounds on the radius overestimation by neglecting the roundoff error. We discuss here several possible measures for interval approximations. We then bound the roundoff error and compare experimentally this bound with the global error distribution on several random data sets. This approach enlightens the relative importance of the roundoff and arithmetic errors depending on the value and homogeneity of relative accuracies of inputs, on the matrix dimension, and on the working precision. This also leads to a new algorithm that is cheaper yet as accurate as previous ones under well-identified conditions.Second, we exploit the parallelism of linear algebra. Previous implementations use calls to BLAS routines on numerical matrices. We show that this may lead to wrong interval results and also restrict the scalability of the performance when the core count increases. To overcome these problems, we implement a blocking version with OpenMP threads executing block kernels with vector instructions. The timings on a 4-octo-core machine show that this implementation is more scalable than the BLAS one and that the cost of numerical and interval matrix products are comparable.

Η παρούσα διδακτορική διατριβή πραγματεύεται το θέμα της αποδοτικής και με βεβαιότητα εύρεσης όλων των ριζών της παραμετρικής εξίσωσης f(x;[p]) = 0, μιας συνεχώς διαφορίσιμης συνάρτησης f με [p] ένα διάνυσμα που περιγράφει όλες τις παραμέτρους της παραμετρικής εξίσωσης και τυποποιούνται με τη μορφή διαστημάτων. Για την επίλυση αυτού του προβλήματος χρησιμοποιήθηκαν εργαλεία της Διαστηματικής Ανάλυσης. Το κίνητρο για την ερευνητική ενασχόληση με το παραπάνω πρόβλημα προέκυψε μέσα από ένα κλασικό πρόβλημα αριθμητικής ανάλυσης: την αριθμητική επίλυση συστημάτων πολυωνυμικών εξισώσεων μέσω διαστηματικής ανάλυσης. Πιο συγκεκριμένα, προτάθηκε μια ευρετική τεχνική αναδιάταξης του αρχικού πολυωνυμικού συστήματος που φαίνεται να βελτιώνει σημαντικά, κάθε φορά, τον χρησιμοποιούμενο επιλυτή. Η ανάπτυξη, καθώς και τα αποτελέσματα αυτής της εργασίας αποτυπώνονται στο Κεφάλαιο 2 της παρούσας διατριβής. Στο επόμενο Κεφάλαιο 3, προτείνεται μια μεθοδολογία για την αποδοτική και αξιόπιστη επίλυση μη-γραμμικών εξισώσεων με διαστηματικές παραμέτρους, δηλαδή την αποδοτική και αξιόπιστη επίλυση διαστηματικών εξισώσεων. Πρώτα, δίνεται μια νέα διατύπωση της Διαστηματικής Αριθμητικής και αποδεικνύεται η ισοδυναμία της με τον κλασσικό ορισμό. Στη συνέχεια, χρησιμοποιείται η νέα διατύπωση της Διαστηματικής Αριθμητικής ως θεωρητικό εργαλείο για την ανάπτυξη μιας επέκτασης της διαστηματικής μεθόδου Newton που δύναται να επιλύσει όχι μόνο κλασικές μη-παραμετρικές μη-γραμμικές εξισώσεις, αλλά και παραμετρικές (διαστηματικές) μη-γραμμικές εξισώσεις. Στο Κεφάλαιο 4 προτείνεται μια νέα προσέγγιση για την αριθμητική επίλυση του προβλήματος της Ολικής Βελτιστοποίησης με περιορισμούς διαστήματα, χρησιμοποιώντας τα αποτελέσματα του Κεφαλαίου 3. Το πρόβλημα της ολικής βελτιστοποίησης, ανάγεται σε πρόβλημα επίλυσης διαστηματικών εξισώσεων, και γίνεται εφικτή η επίλυσή του με τη βοήθεια των θεωρητικών αποτελεσμάτων και της αντίστοιχης μεθοδολογίας του Κεφαλαίου 3. Στο τελευταίο Κεφάλαιο δίνεται μια νέα αλγοριθμική προσέγγιση για το πρόβλημα της επίλυσης διαστηματικών πολυωνυμικών εξισώσεων. Η νέα αυτή προσέγγιση, βασίζεται και γενικεύει την εργασία των Hansen και Walster, οι οποίοι πρότειναν μια μέθοδο για την επίλυση διαστηματικών πολυωνυμικών εξισώσεων 2ου βαθμού. / In this dissertation the problem of finding reliably and with certainty all the zeros a pa-rameterized equation f(x;[p]) = 0, of a continuously differentiable function f is considered, where [p] is an interval vector describing all the parameters of the Equation, which are formed with interval numbers. For this kind of problem, methods of Interval Analysis are used. The incentive to this scientific research was emerged from a classic numerical analysis problem: the numerical solution of polynomial systems of equations using interval analysis. In particular, a heuristic reordering technique of the initial polynomial systems of equations is proposed. This approach seems to improve significantly the used solver. The proposed technique, as well as the results of this publication are presented in Chapter 2 of this dissertation. In the next Chapter 3, a methodology is proposed for solving reliably and efficiently parameterized (interval) equations. Firstly, a new formulation of interval arithmetic is given and the equivalence with the classic one is proved. Then, an extension of interval Newton method is proposed and developed, based on the new formulation of interval arithmetic. The new method is able to solve not only classic non-linear equations but, non-linear parameterized (interval) equation too. In Chapter 4 a new approach on solving the Box-Constrained Global Optimization problem is proposed, based on the results of Chapter 3. In details, the Box-Constrained Global Optimization problem is reduced to a problem of solving interval equations. The solution of this reduction is attainable through the methodology developed in Chapter 3. In the last Chapter of this dissertation a new algorithmic approach is given for the problem of solving reliably and with certainty an interval polynomial equation of degree $n$. This approach consists in a generalization of the work of Hansen and Walster. Hansen and Walster proposed a method for solving only quadratic interval polynomial equations

Model predictive control is a popular research focus in electric motor control as it allows designers to specify optimization goals and exhibits fast transient response. Availability of faster and more affordable computers makes it possible to implement these algorithms in real-time. Real-time implementation is not without challenges however as these algorithms exhibit high computational complexity. Field-programmable gate arrays are a potential solution to the high computational requirements. However, they can be time-consuming to develop for. In this thesis, we present a methodology that reduces the size and development time of field-programmable gate array based fixed-point model predictive motor controllers using automated numerical analysis, optimization and code generation. The methods can be applied to other domains where model predictive control is used. Here, we demonstrate the benefits of our methodology by using it to build a motor controller at various sampling rates for an interior permanent magnet synchronous motor, tested in simulation at up to 125 kHz. Performance is then evaluated on a physical test bench with sampling rates up to 35 kHz, limited by the inverter. Our results show that the low latency achievable in our design allows for the exclusion of delay compensation common in other implementations and that automated reduction of numerical precision can allow the controller design to be compacted. / Thesis / Master of Applied Science (MASc)

科学研究費補助金 研究種目:総合研究(A) 課題番号:04302008 研究代表者:三井 斌友 研究期間:1992-1994年度

L'implantation en Virgule Flottante (VF) d'une fonction à valeurs réelles est réalisée avec arrondi correct si le résultat calculé est toujours égal à l'arrondi de la valeur exacte, ce qui présente de nombreux avantages. Mais pour implanter une fonction avec arrondi correct de manière fiable et efficace, il faut résoudre le «dilemme du fabricant de tables» (TMD en anglais). Deux algorithmes sophistiqués (L et SLZ) ont été conçus pour résoudre ce problème, via des calculs longs et complexes effectués par des implantations largement optimisées. D'où la motivation d'apporter des garanties fortes sur le résultat de ces pré-calculs coûteux. Dans ce but, nous utilisons l'assistant de preuves Coq. Tout d'abord nous développons une bibliothèque d'«approximation polynomiale rigoureuse», permettant de calculer un polynôme d'approximation et un intervalle bornant l'erreur d'approximation à l'intérieur de Coq. Cette formalisation est un élément clé pour valider la première étape de SLZ, ainsi que l'implantation d'une fonction mathématique en général (avec ou sans arrondi correct). Puis nous avons implanté en Coq, formellement prouvé et rendu effectif 3 vérifieurs de certificats, dont la preuve de correction dérive du lemme de Hensel que nous avons formalisé dans les cas univarié et bivarié. En particulier, notre «vérifieur ISValP» est un composant clé pour la certification formelle des résultats générés par SLZ. Ensuite, nous nous sommes intéressés à la preuve mathématique d'algorithmes VF en «précision augmentée» pour la racine carré et la norme euclidienne en 2D. Nous donnons des bornes inférieures fines sur la plus petite distance non nulle entre sqrt(x²+y²) et un midpoint, permettant de résoudre le TMD pour cette fonction bivariée. Enfin, lorsque différentes précisions VF sont disponibles, peut survenir le phénomène de «double-arrondi», qui peut changer le comportement de petits algorithmes usuels en arithmétique. Nous avons prouvé en Coq un ensemble de théorèmes décrivant le comportement de Fast2Sum avec double-arrondis. / The Floating-Point (FP) implementation of a real-valued function is performed with correct rounding if the output is always equal to the rounding of the exact value, which has many advantages. But for implementing a function with correct rounding in a reliable and efficient manner, one has to solve the ``Table Maker's Dilemma'' (TMD). Two sophisticated algorithms (L and SLZ) have been designed to solve this problem, relying on some long and complex calculations that are performed by some heavily-optimized implementations. Hence the motivation to provide strong guarantees on these costly pre-computations. To this end, we use the Coq proof assistant. First, we develop a library of ``Rigorous Polynomial Approximation'', allowing one to compute an approximation polynomial and an interval that bounds the approximation error in Coq. This formalization is a key building block for verifying the first step of SLZ, as well as the implementation of a mathematical function in general (with or without correct rounding). Then we have implemented, formally verified and made effective 3 interrelated certificates checkers in Coq, whose correctness proof derives from Hensel's lemma that we have formalized for both univariate and bivariate cases. In particular, our ``ISValP verifier'' is a key component for formally verifying the results generated by SLZ. Then, we have focused on the mathematical proof of ``augmented-precision'' FP algorithms for the square root and the Euclidean 2D norm. We give some tight lower bounds on the minimum non-zero distance between sqrt(x²+y²) and a midpoint, allowing one to solve the TMD for this bivariate function. Finally, the ``double-rounding'' phenomenon can typically occur when several FP precision are available, and may change the behavior of some usual small FP algorithms. We have formally verified in Coq a set of results describing the behavior of the Fast2Sum algorithm with double-roundings.