An important task in brain modeling is that of estimating model parameters and quantifying their uncertainty. In this thesis we tackle this problem from a Bayesian perspective: we use experimental data to update the prior information about model parameters, in order to obtain their posterior distribution. Uncertainty quantification via a direct computation of the posterior has a prohibitive computational cost in high dimensions. An alternative to a direct computation is offered by Markov chain Monte Carlo (MCMC) methods. The aim of this project is to analyse some of the methods within this class and improve their convergence. In this thesis we describe the following MCMC methods: Metropolis-Hastings (MH) algorithm, Metropolis adjusted Langevin algorithm (MALA), simplified manifold MALA (smMALA) and Approximate Bayesian Computation MCMC (ABCMCMC). SmMALA is further analysed in Paper A, where we propose an algorithm to approximate a key component of this algorithm (the Fisher Information) when applied to ODE models, with the purpose of reducing the computational cost of the method. A theoretical analysis of MCMC methods is carried out in Paper B and relies on tools from the theory of large deviations. In particular, we analyse the convergence of the MH algorithm by stating and proving a large deviation principle (LDP) for the empirical measures produced by the algorithm. Some of the methods analysed in this thesis are implemented in an R package, available on GitHub as “icpm-kth/uqsa” and presented in Paper C, and are applied to subcellular pathway models within neurons in the context of uncertainty quantification of the model parameters. / En viktig uppgift inom hjärnmodellering är att uppskatta parametrar i modellen och kvantifiera deras osäkerhet. I denna avhandling hanterar vi detta problem från ett Bayesianskt perspektiv: vi använder experimentell data för att uppdatera a priori kunskap av modellparametrar, för att erhålla deras posteriori-fördelning. Osäkerhetskvantifiering (UQ) via direkt beräkning av posteriorfördelningen har en hög beräkningskostnad vid höga dimensioner. Ett alternativ till direkt beräkning ges av Markov chain Monte Carlo (MCMC) metoder. Syftet med det här projektet är att analysera några MCMC metoder och förbättra deras konvergens. I denna avhandling beskriver vi följande MCMC algoritmer: “Metropolis-Hastings” (MH), “Metropolis adjusted Langevin” (MALA), “Simplified Manifold MALA” (smMALA) och “Approximate Bayesian Computation MCMC” (ABCMCMC). SmMALA analyseras i artikel A. Där presenterar vi en algoritm för att approximera en nyckelkomponent av denna algoritm (Fisher informationen) när den tillämpas på ODE modeller i syfte att minska metodens beräkningskostnad. En teoretisk analys av MCMC metoder behandlas i artikel B och bygger på verktyg från teorin av stora avvikelser. Mer specifikt, vi analyserar MH algoritmens konvergens genom att formulera och bevisa en stora avvikelser princip (LDP) för de empiriska mått som produceras av algoritmen. Några av metoderna analyserade i den här avhandlingen har implementerats i ett R paket som finns på GitHub som “icpm-kth/uqsa” och presenteras i artikel C. Metoderna tillämpas på subcellulära vägmodeller inom neuroner i sammanhanget av osäkerhetskvantifieringen av modellparametrar. / <p>QC 2023-08-21</p>

[pt] O estudo de desvios, e em particular grandes desvios, tem uma história longa na teoria de probabilidade. Nas últimas décadas muitos artigos consideraram essas questões no contexto de subgrafos de grafos aleatórios G(n, p) e G(n, m). Esta dissertação considera a cauda inferior para o número de triângulos no grafo aleatório G(n, m). Duas abordagens estão consideradas: Martingales, a partir artigo de Christina Goldschmidt, Simon Griffiths e Alex Scott; e Teoria Espectral de Grafos, a partir do artigo de Joe Neeman, Charles Radin e Lorenzo Sadun. Essas duas abordagens conseguem encontrar o comportamento da cauda em dois regimes diferentes. Na dissertação discutiremos a visão geral do artigo de Goldschmidt, Griffiths e Scott, e discutiremos em detalhes o artigo de Neeman, Radin e Sadun. Em particular, exploraremos a conexão entre a cauda inferior do número de triângulos e o comportamento dos autovalores mais negativos da matriz de adjacência. Veremos que a contagem tende a depender, essencialmente, do autovalor mais negativo. / [en] The study of deviations, and in particular large deviations, has a long history in Probability Theory. In recent decades many articles have considered these questions in the context of subgraphs of the random graphs G(n, p) and G(n, m). This dissertation considers the lower tail for the number of triangles in the random graph G(n, m). Two approaches are considered: Martingales, based on the article of Christina Goldschmidt, Simon Griffiths and Alex Scott; and Spectral Graph Theory, based on the article of Joe Neeman, Charles Radin and Lorenzo Sadun. These two approaches manage to find the behavior of the tail in two different regimes. In this dissertation we give an overview of the article of Goldschmidt, Griffiths and Scott, discuss in detail the article of artigo Neeman, Radin and Sadun. In particular, we shall explore the connection between the lower tail of the number of triangles and the behavior of the most negative eigenvalues of the adjacency matrix. We shall see that the triangle count tends to especially depend on the most negative eigenvalue.

Postnord TPL är en ledande aktör inom 3PL-branschen i Norden. På önskemål av företaget har en studie genomförts av returprocessen för Postnord TPLs största kund inom elektronikbranschen. En effektiv hantering av returprocessen är en viktig del för tredjepartslogistikföretag för att behålla konkurrenskraft på marknaden. Den effektiva hanteringen är grunden för att skapa en returprocess med låga kostnader, hög grad av resursutnyttjande och tillfredsställande av kundens önskemål.  Syftet med studien är att identifiera avvikelser i returprocessen och de bakomliggande orsakerna samt att föreslå förbättringsåtgärder för att reducera avvikelser och effektivisera returprocessen vid Postnord TPL. För att uppnå syftet har författarna genomfört en fallstudie av returprocessen för Postnord TPL. Genom deltagande observationer, processpromenad och semistrukturerade intervjuer med anställda har empiri samlats in för att skapa en förståelse av processen. Empirin har analyserats för att identifiera avvikelser och möjliga förbättringsåtgärder. Analysen i studien identifierade ett antal avvikelser i returprocessen och de bakomliggande orsakerna till att de uppstår. Avvikelserna berörde onödiga förflyttningar, överarbetning, mänskliga fel samt kompetens och kunskap hos nuvarande personal. Författarna har via observationerna och intervjuerna identifierat var avvikelserna uppstår. Ett orsak-verkan diagram har tagits fram för att definiera de bakomliggande orsakerna till avvikelserna. Vidare har författarna genom en analys av avvikelserna och orsakerna formulerat konkreta förbättringsförslag för att reducera eller eliminera avvikelserna. Förbättringsåtgärderna som presenteras är; flytt av processen till samma våning, eliminera dubbelhantering av returer från företag, använda samma system, skapa rörlighetsklassificering av produkter, skapa nya standarder och rutiner, arbeta med förebyggande analyser, implementera PDCA-cykeln och Lean-verktyg för arbete med ständig förbättring samt att skapa tydliga mål för processen. Målet med en implementering av förbättringsförslagen är att reducera avvikelser och därigenom bidra till ökad effektivitet och konkurrenskraft för Postnord TPL. Samarbetet med företaget har varit avgörande för att säkerställa relevansen och genomförbarheten av förslagen. / Postnord TPL is a leading company in the 3PL-industry in the Nordics. At the company's request, a study was conducted on the return process for Postnord TPL's largest customer in the electronics industry. Efficient handling of the return process is a crucial aspect for third-party logistics companies to maintain competitiveness in the market. Efficient management forms the foundation for creating a return process with low costs, high resource utilization and customer satisfaction. The purpose of the study is to identify deviations in the return process and the underlying causes, as well as to propose improvement measures to reduce deviations and create a more efficient return process at Postnord TPL. To achieve this goal, the authors conducted a case study of Postnord TPL's return process. Through participatory observations, process walkthroughs, and semi-structured interviews with employees, empirical data was collected to understand the process. The empirical data was analyzed to identify deviations and possible improvement measures. The study's analysis identified several deviations in the return process and the underlying causes of their occurrence. The deviations involved unnecessary movements, overprocessing, human errors, as well as the competence and knowledge of current staff. Through observations and interviews, the authors identified where the deviations occur. A cause-and-effect diagram was created to define the underlying causes of the deviations. Furthermore, through an analysis of the deviations and their causes, the authors formulated concrete improvement suggestions to reduce or eliminate the deviations. The improvement measures presented include; relocating the process to the same floor, eliminating double handling of returns from companies, using the same system, creating a mobility classification of products, establishing new standards and routines, working with preventive analyses, implementing the PDCA cycle and Lean tools for continuous improvement, and setting clear goals for the process. The goal of implementing these improvement suggestions is to reduce deviations and thereby contribute to increased efficiency and competitiveness for Postnord TPL. Collaboration with the company has been crucial to ensure the relevance and feasibility of the proposals.

Diese Arbeit behandelt die Asymptotik von Austritts- und Übergangszeiten für gewisse schwach zeitinhomogene Diffusionsprozesse. Darauf basierend wird ein probabilistischer Begriff der stochastischen Resonanz (SR) studiert. Techniken der großen Abweichungen spielen eine zentrale Rolle. Im ersten Teil der Arbeit (Kapitel 1-3) werden Resultate aus der Theorie der großen Abweichungen für zeithomogene Diffusionen rekapituliert. Es werden die klassischen Resultate von Freidlin und Wentzell und Erweiterungen dieser Theorie präsentiert, und es wird an das Kramers''sche Austrittszeitengesetz erinnert. Teil II befasst sich mit dem Phänomen der SR, d.h. mit Periodizitätseigenschaften von Diffusionen. In Kapitel 4 werden physikalische Maße zur Messung der Periodizität diskutiert. Deren Nachteile legen es nahe, einem alternativen, probabilistischen Ansatz zu folgen, der hier behandelt wird. Das 5. Kapitel dient der Herleitung eines gleichmäßigen Prinzips der großen Abweichungen für Diffusionen mit schwach zeitabhängigem, periodischem Drift. Die Gleichmäßigkeit des Prinzips ermöglicht die exakte Bestimmung exponentieller Übergangsraten in Kapitel 6, das die zentralen Ergebnisse des 2. Teils beinhaltet. Hierdurch wird die Maximierung gewisser Übergangswahrscheinlichkeiten ermöglicht, was zum in Kapitel 7 studierten Resonanzbegriff führt. Teil III der Arbeit setzt sich mit der Asymptotik von Austrittszeiten sogenannter selbststabilisierender Diffusionen auseinander. In Kapitel 8 wird der Zusammenhang zwischen interagierenden Teilchensystemen und selbststabilisierenden Diffusionen erläutert und die Existenz- und Eindeutigkeitsfrage behandelt. Das 9. Kapitel dient dem Studium der großen Abweichungen dieser Klasse von Diffusionen. In Kapitel 10 wird das Kramers''sche Austrittszeitengesetz auf selbststabilisierende Diffusionen übertragen, und in Kapitel 11 wird der Einfluß der selbststabilisierenden Komponente auf das Austrittszeitengesetz illustriert. / In this thesis, we study the asymptotic behavior of exit and transition times of certain weakly time inhomogeneous diffusion processes. Based on these asymptotics, a probabilistic notion of stochastic resonance (SR) is investigated. Large deviations techniques play the key role throughout this work. In the first part (Chapters 1-3) we recall the large deviations theory for time homogeneous diffusions. We present the classical results due to Freidlin and Wentzell and extensions thereof, and we remind of Kramers'' exit time law. Part II deals with the phenomenon of stochastic resonance. That is, we study periodicity properties of diffusion processes. In Chapter 4 we explain the paradigm of stochastic resonance and discuss physical notions of measuring periodicity of diffusions. Their drawbacks suggest to follow an alternative probabilistic approach, which is treated in this work. In Chapter 5 we derive a large deviations principle for diffusions subject to a weakly time dependent periodic drift term. The uniformity of the obtained large deviations bounds w.r.t. the system''s parameters plays a key role for the treatment of transition time asymptotics in Chapter 6, which contains the main result of the second part. The exact exponential transition rates obtained here allow for maximizing transition probabilities, which finally leads to the announced probabilistic notion of resonance studied in Chapter 7. In the third part we investigate the exit time asymptotics of a certain class of so-called self-stabilizing diffusions. In Chapter 8 we explain the connection between interacting particle systems and self-stabilizing diffusions, and we address the question of existence. The following Chapter 9 is devoted to the study of the large deviations behavior of these diffusions. In Chapter 10 Kramers'' exit law is carried over to our class of self-stabilizing diffusions. Finally, the influence of self-stabilization is illustrated in Chapter 11.

Face à la demande de plus en plus exigeante en terme de qualité et de coût de fabrication des produits manufacturés, la qualification et quantification optimal des défauts acceptables est primordial. Le tolérancement est le moyen de communication permettant de définir les variations géométriques autorisé entre les différents corps de métier intervenant au cours du cycle de fabrication du produit. Un tolérancement optimal est le juste compromis entre coût de fabrication et qualité du produit final. Le tolérancement repose sur 3 problématiques majeures: la spécification (normalisation d'un langage complet et univoque), la synthèse et l'analyse de tolérances. Nous proposons dans ce document de nouvelles méthodes d'analyse et de synthèse du tolérancement tridimensionnel. Ces méthodes se basent sur une modélisation de la géométrie à l'aide de l'outil domaine jeux et écarts développé au laboratoire. La première étape consiste à déterminer les différentes topologies composant un mécanisme tridimensionnel. Pour chacune de ces topologies est définie une méthode de résolution des problématiques de tolérancement. Au pire des cas, les conditions de respect des exigences fonctionnelles se traduisent par des conditions d'existence et d'inclusions sur les domaines. Ces équations de domaines peuvent ensuite être traduites sous forme de système d'inéquations scalaires. L'analyse statistique s'appuie sur des tirages de type Monte-Carlo. Les variables aléatoires sont les composantes de petits déplacements des torseur écarts défini à l'intérieur de leur zone de tolérance (modélisée par un domaine écarts) et les dimensions géométriques fixant l'étendue des jeux (taille du domaine jeux associé). A l'issue des simulations statistiques, il est possible d'estimer le risque de non-qualité et les jeux résiduels en fonction du tolérancement défini. Le développement d'une nouvelle représentation des domaines jeux et écarts plus adapté, permet de simplifier les calculs relatifs aux problématiques de tolérancement. Le traitement local de chaque topologie élémentaire de mécanisme permet d'effectuer le traitement global des mécanismes tridimensionnels complexes avec prise en compte des jeux. / As far as the demand in quality and cost of manufacturing increase, the optimal qualification and quantification of acceptable defects is essential. Tolerancing is the means of communication between all actors of manufacturing. An optimal tolerancing is the right compromise between manufacturing cost and quality of the final product. Tolerancing is based on three major issues: The specification (standardization of a complete and unequivocal language), synthesis and analysis of the tolerancing. We suggest in this thesis some new analysis and synthesis of the three-dimensional tolerancing. These methods are based on a geometric model define by the deviations and clearances domains developed on the laboratory. The first step consists in determining the elementary topology that composes a three-dimensional mechanism. For each kind of these topologies one resolution method is defined. In worst case, the condition of functional requirement respect is traduced by existence and inclusions conditions on the domains. Then these domains equations can be translated in inequalities system of scalar. The statistical analysis uses the Monte-Carlo simulation. The random variables are the small displacements components of the deviation torsor which is defined inside its tolerance area (model by a deviations domain) and the geometrics dimensions which set the extent of clearance (size of the clearance domain). Thanks to statistical simulation, it is possible to estimate the non-quality rate in regards to the defined tolerancing. The development of a new representation of clearances and deviations domains most suitable, allows us to simplify the calculation for tolerancing problems. The local treatment of elementary topology makes enables the global treatment of complex three-dimensional mechanisms with take into account of clearances.

The main theme of this master thesis is the design of technological process measurement and processing of measured data entered crane tracks to determine the geometric parameters of these crane tracks. Diploma thesis further describes the procedure for testing and possible rectification surveying equipment and instrumentation. The thesis is divided of 7 parts, contains 31 images, 4 graphs, 44 tables and 13 attachments. The theoretical part of this thesis contains information about several types of cranes, methods of measurement and processing, as well information on legal and technical regulations related to this topic of the thesis. The fourth and fifth parts of this thesis provides information on selected crane tracks on which the measurements were taken, the choice of methods of measurement, processing technique and the results obtained. The final part of the thesis includes evaluation of the results obtained, comparing methods of measurement, processing, evaluation and design of a technique of determining the geometric parameters of crane tracks.

Summaries in Afrikaans and English / The aim of this study is to investigate the developmental deficits among milieu disadvantaged pre-schoolers in a multicultural learning environment and to identify the cause of these deficits. The availability of school readiness programmes and whether these programmes fulfil in the needs of milieu disadvantaged pre-schoolers are investigated. In the light of the theoretical and empirical research it appears that the profile of milieu disadvantaged pre-schoolers is in a process of change. Developmental shortcomings are experienced with regard to the following developmental aspects: Emotional Physical Cognitive Social, moral and aesthetical development According to the theoretical and empirical research these developmental deficits are caused by factors due to the home environment, the school as well as socio-demographic and socio-economic factors. From the research recommendations regarding the following were generated: The parents The pre-primary school The primary school The Department of Education Further researchSee file / Die doel met die onderhawige studie is om te bepaal watter ontwikkelingstekorte by milieubenadeelde kleuters in 'n multikulturele leeromgewing voorkom en om die faktore te identifiseer waardeur hierdie tekorte veroorsaak word. Daar word 'n breedvoerige blik gewerp op beskikbare leergereedheidsprogramme en in hoe 'n mate dit die ontwikkelingstekorte van milieubenadeelde kleuters aanspreek. Aan die hand van die literatuurstudie en 'n empiriese ondersoek blyk dit dat die beeld van milieubenadeelde kleuters besig is om te verander. Ontwikkelingstekorte word veral ten opsigte van die volgende aspekte ervaar: Emosionele Fisieke Kognitiewe Sosiale, morele en estetiese ontwikkeling Uit die literatuurstudie en die empiriese ondersoek blyk dit dat bogenoemde ontwikkelingstekorte veroorsaak word deur huislike, skolastiese, sosio-demografiese en sosio-ekonomiese faktore. Uit die ondersoek word aanbevelings ten opsigte van die volgende gegenereer: Die ouerhuis Die pre-primere skool Die primere skool Department of Education Verdere navorsing / Psychology of Education / M. Ed. (Sieklundige Opvoedkunde)

Le sujet principal de ce mémoire est l'étude de la distribution asymptotique de la fonction f_m qui compte le nombre de diviseurs premiers distincts parmi les nombres premiers $p_1,...,p_m$. Au premier chapitre, nous présentons les sept résultats qui seront démontrés au chapitre 4. Parmi ceux-ci figurent l'analogue du théorème d'Erdos-Kac et un résultat sur les grandes déviations. Au second chapitre, nous définissons les espaces de probabilités qui serviront à calculer les probabilités asymptotiques des événements considérés, et éventuellement à calculer les densités qui leur correspondent. Le troisième chapitre est la partie centrale du mémoire. On y définit la promenade aléatoire qui, une fois normalisée, convergera vers le mouvement brownien. De là, découleront les résultats qui formeront la base des démonstrations de ceux chapitre 1. / The main topic of this masters thesis is the study of the asymptotic distribution of the fonction f_m which counts the number of distinct prime divisors among the first $m$ prime numbers, i.e. $p_1,...,p_m$. The first chapter provides the seven main results which will later on be proved in chapter 4. Among these we find the analogue of the Erdos-Kac central limit theorem and a result on large deviations. In the following chapter, we define several probability spaces on which we will calculate asymptotic probabilities of specific events. These will become necessary for calculating their corresponding densities. The third chapter is the main part of this masters thesis. In it, we introduce a random walk which, when suitably normalized, will converge to the Brownian motion. We will then obtain results which will form the basis of the proofs of those of chapiter 1.

Dans cette thèse on s'intéresse au temps local d'auto-intersections de marches aléatoires. Cette quantité est définie comme la norme-p à la puissance p du temps local de la marche. Elle regarde dans quelle mesure la trajectoire de la marche aléatoire s'intersecte. Le temps local d'auto-intersections est lié à différents modèles physiques comme les modèles de polymères ou les problèmes d'écoulements de flux en milieux stratifiés mais aussi au modèle mathématiques des marches aléatoires en paysages aléatoires. Nous nous sommes pour notre part intéressés en particulier aux grandes déviations du temps local d'auto-intersections, c'est à dire que nous regardons la probabilité que la quantité d'intersections de la marche aléatoire soit plus grande que sa moyenne. Cette question qui a été très étudiée au cours des années 2000 fait apparaitre trois cas distincts, le cas sous-critique, le cas critique et le cas sur-critique. Nous améliorons la connaissance sur cette question au travers de deux résultats complets et d'un résultat partiel. D'abord nous prouvons un principe de grandes déviations dans les cas critique et sur-critique des marches alpha-stables, puis nous améliorons les échelles de déviations au cas sous-critique tout entier de la marche simple, enfin nous sommes en train d'étendre ce dernier résultat aux marches alpha-stables. Par ailleurs les trois preuves sont basées sur l'utilisation d'une version due à Eisenbaum d'un théorème d'isomorphisme de Dynkin. Cette méthode d'abord introduite par Castell dans le cas critique est donc ici étendue aux autres cas. Nous avons donc réussi à unifier les différentes méthodes de preuves au travers ce théorème d'isomorphisme. / In this thesis we are interested in the self-intersection local times of random walks. This quantity is defined as the p-norm to the power of p of the local times of the random walk. It measures how much the trajectory of the random walk intersects itself. The self-intersection local times is connected with various physical models as polymer models or problems of anomalous dispersion in layered random flows, but it is also linked with the mathematical model of random walks in random sceneries. More precisely, we are interested in the large deviations of the self-intersection local times, i.e. we work on the probability for the intersections to be larger than expected. This question that has been studied a lot during the 2000's is divided in three cases, the subcritical one, the critical one and the super critical one. We improve the knowledge about this question by two complete results and a partial one. First, we have proved a large deviation principle in the critical and super critical cases of alpha-stable random walks, then we have improved the deviations' scales to the entire subcritical case of simple random walk, finally we are extending this last result to the alpha-stable random walks. The three proofs are based on a version due to Eisenbaum of a Dynkin isomorphism theorem. This method which has been first introduced by Castell in the critical case, is extended here to the others cases. Thus, we have succeeded to unify the methods of proof by this isomorphism theorem.

Der Beitrag unterteilt sich in zwei Teile: Teil 1 (vgl. Informationen/07; 1976,05) und Teil 2 (cp. Informationen/07; 1976,06). Teil 1 enthält eine Einleitung und Grenzwertsätze für unabhängige und identisch verteilte Zufallsgrößen und die Übertragung der betrachteten Grenzwertsätze auf den Fall der Existenz einseitiger Momente. Teil 2 enthält Grenzwertsätze für mittlere Abweichungen für Summen unabhängiger nichtidentisch verteilter Zufallsgrößen (Serienschema) und eine Diskussion der erhaltenen Ergebnisse und schließlich einige Literaturangaben. Sei F_n(x) die Verteilungsfunktion der Summe X_1+X_2+...+X_n, wobei X_1, X_2, ...,X_n unabhängige und identisch verteilte Zufallsgrößen mit Erwartungswert 0 und Streuung 1 und endlichen absoluten Momenten c_m, m>2, sind, und sei Phi die standardisierte Normalverteilungsfunktion. Es werden absolute Konstanten L_i derart berechnet, dass wir Fehlerabschätzungen im unleichmäßigen zentralen Grenzwertsätzen in verschiedenen Fällen angeben können, wobei sich der Index i in L_i auf folgende fünf Fälle bezieht: kleine x, mittlere Abweichungen für x, große Abweichungen für x, kleine n und große n. Im Fall der Existenz einseitiger Momente werden obere Schanken für 1-F_n(x) angegeben für x>D_m*n^(1/2)*ln(n) bzw. x>D_m*n^(1/2)*(ln(n))^(1/2), womit Ergebnisse von S.V.NAGAEV(1965) präzisiert werden. / The paper is divided in two parts: part 1 (cp. Informationen/07; 1976,05) and part 2 (cp. Informationen/07; 1976,06). Part 1 contains an introduction and limit theorems for iid random variables and the transfer of the considered limit theorems to the case of the existence of onesided moments. Part 2 contains limit theorems of moderate deviations for sums of series of non iid random variables and a discussion of all obtained results in part 1 and 2 and finally some references. Let F_n(x) be the cdf of X_1+X_2+...+X_n, where X_1, X_2, ...,X_n are iid random variables with mean 0 and variance 1 and with m-th absolute moment c_m, m>2, and Phi the cdf of the unit normal law. Explicit universal constants L_i are computed such that we have an error estimate in the nonuniform central limit theorem with the L_i, where i corresponds to the five cases considered: small x, moderate deviations for x, large deviations for x, small n , large n. Additional upper bounds for 1-F_n(x) are obtained if the one-sided moments of order m, m>2, are finite and if x>D_m*n^(1/2)*ln(n) and x>D_m*n^(1/2)*(ln(n))^(1/2) respectively improving results by S.V.NAGAEV (1965).