Alors que la complexité des fonctions d'ordre 1 est bien définie et étudiée, il n'existe pas de notion satisfaisante à tout ordre. Une telle théorie existe déjà à l'ordre 2 et permet de définir une classe analogue aux fonctions calculables en temps polynomial usuelles. Cela est tout particulièrement intéressant dans le domaine de l'analyse récursive où l'on peut représenter, entre autres, les nombres et les fonctions réelles par des fonctions d'ordre 1. On peut alors remarquer un lien fort entre calculabilité et continuité, et aussi rapprocher la complexité avec certaines propriétés analytiques, ce que nous illustrons dans le cas des opérateurs réels. Nous prouvons cependant que, du point de vue de la complexité, les fonctions d'ordre 1 ne permettent pas de représenter fidèlement certains espaces mathématiques. Ce résultat appuie tout particulièrement la nécessité d'une théorie de la complexité d'ordre supérieur. Nous développons alors un modèle de calcul basé sur la sémantique des jeux, où l'application de deux fonctions est représentée par la confrontation de deux stratégies dans un jeu. En définissant la taille de telles stratégies, nous pouvons déduire une notion robuste et pertinente de complexité pour ces stratégies et donc pour les fonctions d'ordre supérieur. Nous définissons aussi une classe de fonctions calculables en temps polynomial qui paraît être un bon candidat pour définir une classe de complexité raisonnable à tout ordre / While first order complexity is well defined and studied, higher order lacks a satisfactory notion of complexity. Such a theory already exists at order 2 and provides a complexity class analogue to usual polynomial time computable functions. This is already especially interesting in the domain of computable analysis, where real numbers or real functions for example can be represented by first order functions. In particular, there is a clear link between computability and continuity, and we also illustrate in the case of real operators that complexity can be related to some analytical properties. However, we prove that, from a complexity point of view, some mathematical spaces can not be faithfully represented by order 1 functions and require higher order ones. This result underlines that there is a need to develop a notion of complexity at higher types which will be, in particular but not only, useful to computable analysis. We have developed a computational model for higher type sequential computations based on a game semantics approach, where the application of two functions is represented by the game opposing two strategies. By defining the size of such strategies, we are able to define a robust and meaningful notion of complexity at all types, together with a class of polynomial time computable higher order functionals which seems to be a good candidate for a class of feasible functionals at higher types

<p>In this thesis we introduce and study a notion of effectivity (or computability) for test functions and for distributions. This is done using the theory of effective (Scott-Ershov) domains and effective domain representations.</p><p>To be able to construct effective domain representations of the spaces of test functions considered in distribution theory we need to develop the theory of admissible domain representations over countable pseudobases. This is done in the first paper of the thesis. To construct an effective domain representation of the space of distributions, we introduce and develop a notion of partial continuous function on domains. This is done in the second paper of the thesis. In the third paper we apply the results from the first two papers to develop an effective theory of distributions using effective domains. We prove that the vector space operations on each space, as well as the standard embeddings into the space of distributions effectivise. We also prove that the Fourier transform (as well as its inverse) on the space of tempered distributions is effective. Finally, we show how to use convolution to compute primitives on the space of distributions. In the last paper we investigate the effective properties of a structure theorem for the space of distributions with compact support. We show that each of the four characterisations of the class of compactly supported distributions in the structure theorem gives rise to an effective domain representation of the space. We then use effective reductions (and Turing-reductions) to study the reducibility properties of these four representations. We prove that three of the four representations are effectively equivalent, and furthermore, that all four representations are Turing-equivalent. Finally, we consider a similar structure theorem for the space of distributions supported at 0.</p>

In this thesis we introduce and study a notion of effectivity (or computability) for test functions and for distributions. This is done using the theory of effective (Scott-Ershov) domains and effective domain representations. To be able to construct effective domain representations of the spaces of test functions considered in distribution theory we need to develop the theory of admissible domain representations over countable pseudobases. This is done in the first paper of the thesis. To construct an effective domain representation of the space of distributions, we introduce and develop a notion of partial continuous function on domains. This is done in the second paper of the thesis. In the third paper we apply the results from the first two papers to develop an effective theory of distributions using effective domains. We prove that the vector space operations on each space, as well as the standard embeddings into the space of distributions effectivise. We also prove that the Fourier transform (as well as its inverse) on the space of tempered distributions is effective. Finally, we show how to use convolution to compute primitives on the space of distributions. In the last paper we investigate the effective properties of a structure theorem for the space of distributions with compact support. We show that each of the four characterisations of the class of compactly supported distributions in the structure theorem gives rise to an effective domain representation of the space. We then use effective reductions (and Turing-reductions) to study the reducibility properties of these four representations. We prove that three of the four representations are effectively equivalent, and furthermore, that all four representations are Turing-equivalent. Finally, we consider a similar structure theorem for the space of distributions supported at 0.

京都大学 / 0048 / 新制・課程博士 / 博士(理学) / 甲第20157号 / 理博第4242号 / 新制||理||1610(附属図書館) / 京都大学大学院理学研究科数学・数理解析専攻 / (主査)准教授 照井 一成, 教授 岡本 久, 教授 長谷川 真人 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DFAM

<p>The thesis consists of three research projects concerning mathematical models for analog computers, originally developed by John Tucker and Jeff Zucker. The models are capable of representing systems that essentially “diverge,” exhibiting no valid behaviour---much the way that digital computers are capable of running programs that never halt. While there is no solution to the general Halting Problem, there are certainly theorems that identify large collections of instances that are guaranteed to halt. For example, if we use a simplified language featuring only assignment, branching, algebraic operations, and loops whose bounds must be fixed in advance (i.e. at “compile time”), we know that all instances expressible in this language will halt.</p> <p>In this spirit, one of the major objectives of all three thesis projects is identify a large class of instances of analog computation (analog computer + input) that are guaranteed to “converge.” In our semantic models, this convergence is assured if a certain operator (representing the computer and its input) has a unique fixed point. The first project is based on an original fixed point construction, while the second and third projects are based on Tucker and Zucker's construction. The second project narrows the scope of the model to a special case in order to concretely identify a class of operators with well-behaved fixed points, and considers some applications. The third project goes the opposite way: widening the scope of the model in order to generalize it.</p> / Doctor of Philosophy (PhD)

Markov partitions provide a 'good' mechanism of symbolic dynamics for uniformly hyperbolic systems, forming the classical foundation for the thermodynamic formalism in this setting, and remaining useful in the modern theory. Usually, however, one takes Bowen's 1970's general construction for granted, or restricts to cases with simpler geometry (as on surfaces) or more algebraic structure. This thesis examines several questions on the algorithmic content of (topological) Markov partitions, starting with the pointwise, entropy-like, topological conjugacy invariant known as orbit complexity. The relation between orbit complexity de nitions of Brudno and Galatolo is examined in general compact spaces, and used in Theorem 2.0.9 to bound the decrease in some of these quantities under semiconjugacy. A corollary, and a pointwise analogue of facts about metric entropy, is that any Markov partition produces symbolic dynamics matching the original orbit complexity at each point. A Lebesgue-typical value for orbit complexity near a hyperbolic attractor is also established (with some use of Brin-Katok local entropy), and is technically distinct from typicality statements discussed by Galatolo, Bonanno and their co-authors. Both our results are proved adapting classical arguments of Bowen for entropy. Chapters 3 and onwards consider the axiomatisation and computable construction of Markov partitions. We propose a framework of 'abstract local product structures'

L'objet de cette thèse est l'étude de l'auto-organisation dans les automates cellulaires unidimensionnels.Les automates cellulaires sont un système dynamique discret ainsi qu'un modèle de calcul massivement parallèle, ces deux aspects s'influençant mutuellement. L'auto-organisation est un phénomène où un comportement organisé est observé asymptotiquement, indépendamment de la configuration initiale. Typiquement, nous considérons que le point initial est tiré aléatoirement: étant donnée une mesure de probabilité décrivant une distribution de configurations initiales, nous étudions son évolution sous l'action de l'automate, le comportement asymptotique étant décrit par la(les) mesure(s) limite(s).Notre étude présente deux aspects. D'abord, nous caractérisons les mesures qui peuvent être atteintes à la limite par les automates cellulaires; ceci correspond aux différents comportements asymptotiques pouvant apparaître en simulation. Cette approche rejoint divers résultats récents caractérisant des paramètres de systèmes dynamiques par des conditions de calculabilité, utilisant des outils d'analyse calculable. Il s'agit également d'une description de la puissance de calcul des automates cellulaires sur les mesures.Ensuite, nous proposons des outils pour létude de l'auto-organisation dans des classes restreintes. Nous introduisons un cadre d'étude d'automates pouvant être vus comme un ensemble de particules en interaction, afin d'en déduire des propriétés sur leur comportement asymptotique. Une dernière direction de recherche concerne les automates convergeant vers la mesure uniforme sur une large classe de mesures initiales (phénomène de randomisation). / The subject of this thesis is the study of self-organization in one-dimensional cellular automata.Cellular automata are a discrete dynamical system as well as a massively parallel model of computation, both theseaspects influencing each other. Self-organisation is a phenomenon where an organised behaviour is observed asymptotically, regardless of the initial configuration. Typically, we consider that the initial point is sampled at random; that is, we consider a probability measure describing the distribution of theinitial configurations, and we study its evolution under the action of the automaton, the asymptoticbehaviour being described by the limit measure(s).Our work is two-sided. On the one hand, we characterise measures that can bereached as limit measures by cellular automata; this corresponds to the possible kinds of asymptoticbehaviours that can arise in simulations. This approach is similar to several recent results characterising someparameters of dynamical systems by computability conditions, using tools from computable analysis. Thisresult is also a description of the measure-theoretical computational power of cellular automata.On the other hand, we provided tools for the practical study of self-organization in restricted classes of cellularautomata. We introduced a frameworkfor cellular automata that can be seen as a set of interacting particles, in order todeduce properties concerning their asymptotic behaviour. Another ongoing research direction focus on cellular automata that converge to the uniform measurefor a wide class of initial measures (randomization phenomenon).