In this paper introduce new approach for unified theory for continuous and discrete time (optimal) control problems based on the generalized Cayley transformation. We also relate the associated discrete and continuous generalized algebraic Riccati equations. We demonstrate the potential of this new approach proving new result for discrete algebraic Riccati equations. But we also discuss where this new approach as well as all other approaches still is non-satisfactory. We explain a discrepancy observed between the discrete and continuous cse and show that this discrepancy is partly due to the consideration of the wrong analogues. We also present an idea for a metatheorem that relates general theorems for discrete and continuous control problems.

No description available.

[pt] A Geometria Diferencial Discreta tem por objetivo desenvolver uma teoria discreta que respeite os aspectos fundamentais da teoria suave. Com isto em mente, são apresentados incialmente resultados da teoria suave da Geometria Afim que terão suas versões discretas tratadas a posteriori. O primeiro objetivo deste trabalho é construir uma estrutura afim discreta para as redes assintóticas definidas no espaço tridimensional, com métrica de Blaschke indefinida e parâmetros assintóticos. Com este intuito, são definidos um campo conormal, que satisfaz as equações de Lelieuvre e está associado a um parâmetro real, e um normal afim que define a forma cúbica da rede e torna a estrutura bem definida. Esta estrutura permite, por exemplo, o estudo das superfícies regradas, com ênfase nas esferas afins impróprias. Além disso, propõe-se uma definição para as singularidades no caso das esferas afins impróprias discretas a partir da construção centrocorda. Outro objetivo deste trabalho é propor uma definição para as superfícies afins discretas com curvatura afim média constante (CAMC), de forma que englobe as superfícies afins mínimas e as esferas afins. As superfícies afins mínimas discretas recebem uma caracterização geométrica bastante interessane e ligada diretamente às quádricas de Lie discretas. O trabalho se completa com o principal resultado, referente à versão discreta das superfícies de Cayley, esferas afins impróprias regradas caracterizadas a partir da conexão afim induzida: uma rede assintótica com CAMC é congruente equiafim à uma superfície de Cayley se, e somente se, a forma cúbica é não nula e a conexão afim induzida é paralela. / [en] Discrete Differential Geometry aims to develop a discrete theory which respects fundamental aspects of smooth theory. With this in mind, some results of smooth theory of Affine Geometry are firstly introduced since their discrete counterparts shall be treated a posteriori. The first goal of this work is construct a discrete affine structure for nets in a three-dimensional space with indefinite Blaschke metric and asymptotic parameters. For this purpose, one defines a conormal vector field, which satisfies Lelieuvre s equations and it is associated to a real parameter; and an affine normal vector field, which defines the cubic form of the net and makes the structure well defined. This structure allows to study, e.g., ruled surfaces with emphasis on improper affine spheres. Moreover, a definition for singularities is proposed in the case of discrete improper affine spheres from the center-chord construction. Another goal here is to propose a definition for an asymptotic net with constant affine mean curvature (CAMC), in a way that encompasses discrete affine minimal surfaces and discrete affine spheres. Discrete affine minimal surfaces receive a beautiful geometrical characterization directly linked to discrete Lie quadrics. This work is completed with the main result about a discrete version of Cayley surfaces, which are ruled improper affine spheres that can be characterized by the induced connection as: an asymptotic net with CAMC is equiaffinely congruent to a Cayley surface if and only if the cubic form does not vanish and the affine induced connection is parallel.

On s'intéresse ici aux actions (discrètes, par isométries) d'un groupe $\Gamma$ sur un espace métrique mesuré $X$ et à la manière dont ces actions écartent les points. Le lemme de Margulis classique conclut lorsque $X$ est une variété simplement connexe de courbure strictement négative et bornée. Une version récente (due à G. Besson, G. Courtois et S. Gallot) conclut lorsque $X$ est un espace métrique mesuré d'entropie bornée, mais est essentiellement limitée au cas où $\Gamma$ est un groupe fondamental d'une variété de courbure négative<br />majorée et de rayon d'injectivité minoré. Nous montrons que ce dernier résultat (et ses applications géométriques) se généralise à une classe ${\cal C}$ plus vaste de groupes (qui contient les groupes hyperboliques selon Gromov, les produits libres et les produits amalgamés ``malnormaux'') et aux quasi-actions par quasi-isométries (avec points fixes éventuels) de ces groupes sur un espace métrique mesuré d'entropie bornée. Nous montrons aussi que ${\cal C}$ est fermé pour une topologie naturelle. Nous appliquons ce résultat au cas où $X$ est le graphe de Cayley d'un groupe $G$ commensurable à un groupe $\Gamma \in {\cal C}$, obtenant des résultats<br />de finitude qui s'appliquent en particulier aux groupes hyperboliques selon Gromov et aux groupes fondamentaux de variétés de diamètre borné. Ces derniers résultats apportent un éclairage nouveau aux questions de l'existence d'un minorant universel de l'entropie pour l'ensemble des groupes $G$ de ce type et de l'existence, pour chacun de ces groupes, d'un système générateur d'entropie algébrique minimale.

It is well known that communication systems employing multiple transmit and multiple receive antennas provide high data rates along with increased reliability. Some of the design criteria of the space-time block codes (STBCs) for multiple input multiple output (MIMO)communication system are that these codes should attain large transmit diversity, high data-rate, low decoding-complexity, low decoding –delay and low peak-to-average power ratio (PAPR). STBCs based on real orthogonal designs (RODs) and complex orthogonal designs (CODs) achieve full transmit diversity and in addition, these codes are single-symbol maximum-likelihood (ML) decodable. It has been observed that the data-rate (in number of information symbols per channel use) of the square CODs falls exponentially with increase in number of antennas and it has led to the construction of rectangular CODs with high rate. We have constructed a class of maximal-rate CODs for n transmit antennas with rate if n is even and if n is odd. The novelty of the above construction is that they 2n+1 are constructed from square CODs. Though these codes have a high rate, this is achieved at the expense of large decoding delay especially when the number of antennas is 5or more. Moreover the rate also converges to half as the number of transmit antennas increases. We give a construction of rate-1/2 CODs with a substantial reduction in decoding delay when compared with the maximal- rate codes. Though there is a significant improvement in the rate of the codes mentioned above when compared with square CODs for the same number of antennas, the decoding delay of these codes is still considerably high. For certain applications, it is desirable to construct codes which are balanced with respect to both rate and decoding delay. To this end, we have constructed high rate and low decoding-delay RODs and CODs from Cayley-Dickson Algebra. Apart from the rate and decoding delay of orthogonal designs, peak-to-average power ratio (PAPR) of STBC is very important from implementation point of view. The standard constructions of square complex orthogonal designs contain a large number of zeros in the matrix result in gin high PAPR. We have given a construction for square complex orthogonal designs with lesser number of zero entries than the known constructions. When a + 1 is a power of 2, we get codes with no zero entries. Further more, we get complex orthogonal designs with no zero entry for any power of 2 antennas by introducing co- ordinate interleaved variables in the design matrix. These codes have significant advantage over the existing codes in term of PAPR. The only sacrifice that is made in the construction of these codes is that the signaling complexity (of these codes) is marginally greater than the existing codes (with zero entries) for some of the entries in the matrix consist of co-ordinate interleaved variables. Also a class of maximal-rate CODs (For mathematical equations pl see the pdf file)

Made available in DSpace on 2014-06-11T19:24:52Z (GMT). No. of bitstreams: 0 Previous issue date: 2014-02-25Bitstream added on 2014-06-13T19:52:43Z : No. of bitstreams: 1 000755405.pdf: 721229 bytes, checksum: a665c9ee190d3a2675b924dd4bb2c525 (MD5) / As origens da Teoria dos Invariantes na Inglaterra e o Mécanique Analytique de Lagrange (1788), é um trabalho voltado principalmente a entender uma possível influência que levou George Boole em 1841, a escrever o artigo Exposition of a General Theory of Linear Transformations e verificar se a motivação que o fez produzir este trabalho é igual ou diferente da motivação que ele exerceu sobre Arthur Cayley e consequentemente sobre James Joseph Sylvester. O presente trabalho apresenta um estudo das origens da Teoria dos Invariantes, no século XIX na Inglaterra. De acordo com os historiadores da Matemática o marco do início desta Teoria foi a publicação de George Boole em 1841. Assumimos este artigo como referência principal para realizar nossa pesquisa. Analisamos “antes” e “após” esta publicação de 1841. Concluímos que o Mécanique Analytique de Lagrange, foi a principal motivação para George Boole escrever seu trabalho e, certamente, George Boole foi uma grande influência para Arthur Cayley no que condiz com a escolha do assunto “invariantes” bem como o desenvolvimento desta Teoria por Cayley / The origins of the theory of invariants in England and Mécanique Analytique of Lagrange (1788), is a work geared primarily to understand a possible influence that led George Boole in 1841, writing the article Exposition of the General Theory of Linear Transformations and verify that the motivation that did produce this work is equal or different of the motivation that he exerted on Arthur Cayley and James Joseph Sylvester consequently. This paper presents a study of the Invariant Theory origins, in the nineteenth century in England. According to historians of Mathematics the beginning of this Theory was the publication in 1841 of George Boole. We have taken this article as a reference to our research. We have proposed to analyzed before and after this publication, 1841. We conclude that the Mécanique Analytique Lagrange, was the essential motivation for George Boole write his work, and certainly George Boole was a great influence to Arthur Cayley in which matches the choice of subject invariants as well as the development of this Theory by Cayley

L'histoire du théorème de Jordan est abordée sous l'angle d'une question d'identité posée sur la période qui sépare la date de 1870 et l'énoncé par Camille Jordan d'une forme canonique des substitutions linéaires des années trente du vingtième siècle au cours desquelles le théorème de Jordan de la décomposition matricielle acquiert une place centrale dans la théorie des matrices canoniques. A partir d'un moment historique de référence, la controverse entre Jordan et Kronecker de 1874, le théorème de Jordan permet de jeter un regard original sur l'histoire de la période 1870-1930 en suivant le rôle joué par des savoirs tacites, des idéaux et des pratiques propres à des réseaux et des communautés. Ce regard permet notamment de mettre en évidence la dynamique d'une tension entre formes canoniques et invariants dans l'évolution de la signification de la notion de forme en mathématiques et contribue à l'histoire de l'algèbre linéaire en décrivant le rôle joué par une méthode de décomposition indissociable d'un mode particulier de représentation : la décomposition matricielle.

In this thesis, the main objects of study are probability measures on the isomorphism classes of countable, connected rooted graphs. An important class of such measures is formed by unimodular measures, which satisfy a certain equation, sometimes referred to as the intrinsic mass transport principle. The so-called law of a finite graph is an example of a unimodular measure. We say that a measure is sustained by a countable graph if the set of rooted connected components of the graph has full measure. We demonstrate several new results involving sustained unimodular measures, and provide thorough arguments for known ones. In particular, we give a criterion for unimodularity on connected graphs, deduce that connected graphs sustain at most one unimodular measure, and prove that unimodular measures sustained by disconnected graphs are convex combinations. Furthermore, we discuss weak limits of laws of finite graphs, and construct counterexamples to seemingly reasonable conjectures.

In this thesis, the main objects of study are probability measures on the isomorphism classes of countable, connected rooted graphs. An important class of such measures is formed by unimodular measures, which satisfy a certain equation, sometimes referred to as the intrinsic mass transport principle. The so-called law of a finite graph is an example of a unimodular measure. We say that a measure is sustained by a countable graph if the set of rooted connected components of the graph has full measure. We demonstrate several new results involving sustained unimodular measures, and provide thorough arguments for known ones. In particular, we give a criterion for unimodularity on connected graphs, deduce that connected graphs sustain at most one unimodular measure, and prove that unimodular measures sustained by disconnected graphs are convex combinations. Furthermore, we discuss weak limits of laws of finite graphs, and construct counterexamples to seemingly reasonable conjectures.