Dynamic Programming Approaches for Estimating and Applying Large-scale Discrete Choice Models

People go through their life making all kinds of decisions, and some of these decisions affect their demand for transportation, for example, their choices of where to live and where to work, how and when to travel and which route to take. Transport related choices are typically time dependent and characterized by large number of alternatives that can be spatially correlated. This thesis deals with models that can be used to analyze and predict discrete choices in large-scale networks. The proposed models and methods are highly relevant for, but not limited to, transport applications. We model decisions as sequences of choices within the dynamic discrete choice framework, also known as parametric Markov decision processes. Such models are known to be difficult to estimate and to apply to make predictions because dynamic programming problems need to be solved in order to compute choice probabilities. In this thesis we show that it is possible to explore the network structure and the flexibility of dynamic programming so that the dynamic discrete choice modeling approach is not only useful to model time dependent choices, but also makes it easier to model large-scale static choices. The thesis consists of seven articles containing a number of models and methods for estimating, applying and testing large-scale discrete choice models. In the following we group the contributions under three themes: route choice modeling, large-scale multivariate extreme value (MEV) model estimation and nonlinear optimization algorithms. Five articles are related to route choice modeling. We propose different dynamic discrete choice models that allow paths to be correlated based on the MEV and mixed logit models. The resulting route choice models become expensive to estimate and we deal with this challenge by proposing innovative methods that allow to reduce the estimation cost. For example, we propose a decomposition method that not only opens up for possibility of mixing, but also speeds up the estimation for simple logit models, which has implications also for traffic simulation. Moreover, we compare the utility maximization and regret minimization decision rules, and we propose a misspecification test for logit-based route choice models. The second theme is related to the estimation of static discrete choice models with large choice sets. We establish that a class of MEV models can be reformulated as dynamic discrete choice models on the networks of correlation structures. These dynamic models can then be estimated quickly using dynamic programming techniques and an efficient nonlinear optimization algorithm. Finally, the third theme focuses on structured quasi-Newton techniques for estimating discrete choice models by maximum likelihood. We examine and adapt switching methods that can be easily integrated into usual optimization algorithms (line search and trust region) to accelerate the estimation process. The proposed dynamic discrete choice models and estimation methods can be used in various discrete choice applications. In the area of big data analytics, models that can deal with large choice sets and sequential choices are important. Our research can therefore be of interest in various demand analysis applications (predictive analytics) or can be integrated with optimization models (prescriptive analytics). Furthermore, our studies indicate the potential of dynamic programming techniques in this context, even for static models, which opens up a variety of future research directions.

Fish nodavirus lytic cycle and semipermissive expression in mammalian and fish cell cultures

In this study, Dicentrarchus labrax encephalitis virus (DIEV), which causes sea bass encephalitis, was propagated in cell culture, thus allowing study of its lytic cycle, DIEV infection of mammalian and fish cells induced different patterns of expression of capsid proteins, which were assembled as virus-like particles, accumulating in the cytoplasm either as diffuse masses or in vesicles, as shown by electron microscopy, These particles correspond to virions, as shown by their ability to induce Secondary infection, Fish cells proved to be more permissive for DIEV than mammalian cells, although virus yield remained low, RNA analysis of infected sea bass cells revealed DIEV RNA3, in addition to genomic RNA1 and RNA2, and the presence of the RNA;! minus strand, thus demonstrating the replication of the DIEV genome, In addition, DIEV RNA-dependent RNA polymerase was associated with mature virions even after purification by a CsCl gradient, but it was dissociated when capsids were destabilized, In addition to providing more information about the relatedness of DIEV to the members of the family Nodaviridae, this study shows that fish nodaviruses may not be able to infect as wide a variety of cells as insect nodaviruses can.

Evaluation des estimations indirectes de mortalité dans trois observatoires de population au Sénégal

En Afrique sub-saharienne, les recensements constituent une source de données fondamentale pour estimer la mortalité, en raison de l’incomplétude de l’enregistrement des décès à l'état civil. Afin d’évaluer la fiabilité des niveaux de mortalité qui peuvent en être déduits, nous appliquons différentes méthodes d’estimation à des extraits des recensements sénégalais de 2002 et 2013, couvrant trois observatoires de population situés à Bandafassi, Mlomp et Niakhar. Les taux de mortalité des moins de cinq ans tirés du nombre d’enfants nés vivants et survivants s’avèrent plus bas que les niveaux attendus au vu des données du suivi démographique. Les estimations déduites des déclarations sur la survie des parents sont largement inférieures aux niveaux de mortalité adulte réellement observés selon le suivi. Par contre, les taux de mortalité par âge basés sur les décès récents déclarés dans les ménages sont conformes aux données du suivi, sauf pour la mortalité infantile qui est sensiblement sous-déclarée en 2002. Cette évaluation confirme que les méthodes indirectes procurent des estimations qui ne peuvent être considérées isolément mais doivent plutôt être systématiquement comparées les unes aux autres. Des études d’évaluation directes menées au niveau individuel à l’aide d’appariements sont nécessaires pour mieux identifier les différentes sources de biais.

Stimulation of bacterial exoproteolytic activity by fish farming in coastal marine ponds: Effect on dissolved protein cycling

The effect of fish farming on dissolved amino acid concentrations, bacterioplankton abundance and exoproteolytic activity was assessed in 3 experimental marine ponds. Different standing stocks of fish were introduced (semi-intensive pond: 250 g.m(-2); semi-extensive pond: 50 g.m(-2) control pond: 0). Sea bass farming increased dissolved combined amino acid (DCAA) concentrations only in the semi-intensive pond. Bacterial standing stock was unaffected by fish food supply. However, bacterial exoproteolytic activity was strongly stimulated by aquaculture intensification; the average maximal rate of dissolved protein hydrolysis (V-m) increased with intensity (control pond: 1 500 nM.h(-1); semi-extensive pond: 2 600 nM.h(-1) semi-intensive pond: 5 100 nM.h(-1)). DCAA fluxes through bacterial exoproteolytic activity ranged between 16 (semi-extensive) and 11% (semi-intensive) of the daily nitrogen input by fish food. Bacterial exoproteolytic activity allowed a substantial part of the increased supply of dissolved amino nitrogen to be incorporated into bacterial biomass, then available for transfer to higher trophic levels within the ponds. It also significantly decreased dissolved organic nitrogen export from the ponds to the surrounding environment.

Traitement des émotions évoquées par les expressions faciales, vocales et musicales à la suite d’un traumatisme craniocérébral

Le traitement des émotions joue un rôle essentiel dans les relations interpersonnelles. Des déficits dans la reconnaissance des émotions évoquées par les expressions faciales et vocales ont été démontrés à la suite d’un traumatisme craniocérébral (TCC). Toutefois, la majorité des études n’ont pas différencié les participants selon le niveau de gravité du TCC et n’ont pas évalué certains préalables essentiels au traitement émotionnel, tels que la capacité à percevoir les caractéristiques faciales et vocales, et par le fait même, la capacité à y porter attention. Aucune étude ne s’est intéressée au traitement des émotions évoquées par les expressions musicales, alors que la musique est utilisée comme méthode d’intervention afin de répondre à des besoins de prise en charge comportementale, cognitive ou affective chez des personnes présentant des atteintes neurologiques. Ainsi, on ignore si les effets positifs de l’intervention musicale sont basés sur la préservation de la reconnaissance de certaines catégories d’émotions évoquées par les expressions musicales à la suite d’un TCC. La première étude de cette thèse a évalué la reconnaissance des émotions de base (joie, tristesse, peur) évoquées par les expressions faciales, vocales et musicales chez quarante et un adultes (10 TCC modéré-sévère, 9 TCC léger complexe, 11 TCC léger simple et 11 témoins), à partir de tâches expérimentales et de tâches perceptuelles contrôles. Les résultats suggèrent un déficit de la reconnaissance de la peur évoquée par les expressions faciales à la suite d’un TCC modéré-sévère et d’un TCC léger complexe, comparativement aux personnes avec un TCC léger simple et sans TCC. Le déficit n’est pas expliqué par un trouble perceptuel sous-jacent. Les résultats montrent de plus une préservation de la reconnaissance des émotions évoquées par les expressions vocales et musicales à la suite d’un TCC, indépendamment du niveau de gravité. Enfin, malgré une dissociation observée entre les performances aux tâches de reconnaissance des émotions évoquées par les modalités visuelle et auditive, aucune corrélation n’a été trouvée entre les expressions vocales et musicales. La deuxième étude a mesuré les ondes cérébrales précoces (N1, N170) et plus tardives (N2) de vingt-cinq adultes (10 TCC léger simple, 1 TCC léger complexe, 3 TCC modéré-sévère et 11 témoins), pendant la présentation d’expressions faciales évoquant la peur, la neutralité et la joie. Les résultats suggèrent des altérations dans le traitement attentionnel précoce à la suite d’un TCC, qui amenuisent le traitement ultérieur de la peur évoquée par les expressions faciales. En somme, les conclusions de cette thèse affinent notre compréhension du traitement des émotions évoquées par les expressions faciales, vocales et musicales à la suite d’un TCC selon le niveau de gravité. Les résultats permettent également de mieux saisir les origines des déficits du traitement des émotions évoquées par les expressions faciales à la suite d’un TCC, lesquels semblent secondaires à des altérations attentionnelles précoces. Cette thèse pourrait contribuer au développement éventuel d’interventions axées sur les émotions à la suite d’un TCC.

Música de câmara de Rodil, Pla e Avondano: a flauta na 2ª metade do séc. XVIII e o estilo galante ibérico

O presente trabalho é uma abordagem teórica e interpretativa à música para traverso de António Rodil (c.1730-1787), Pedro Avondano (1714-1782) e dos irmãos Pla. A interação entre as duas componentes do discurso são fundamentos essenciais para uma global compreensão das obras, sendo por sua vez, foco central no tratamento do estilo galante ibérico. A dissertação esta dividida em 5 capítulos; 1. contextualização da prática musical em Portugal na segunda metade do séc. XVIII; 2. o percurso artístico do flautista António Rodil (c.1730-1787); 3. a música de câmara de Pedro Avondano (1714-1782) e a primeira obra conhecida para o traverso em Portugal A Collection ofLisbon Minuets for two Violins or two German flutes and bass (1766); 4. o traverso na obra dos irmãos PLA; 5. reflexões em torno das questões interpretativas.

The twentieth century jazz piano trio: the rise of an iconic jazz paradigm

Considerando a gravação de 1947 The Bud Powell Trio como a gravação referência de trio de piano de jazz moderno, esta tese centra-se no surgimento e evolução do trio de jazz moderno cujo líder é pianista. Começando por apresentar, uma resenha dos estilos e técnicas para piano de da época pre-Powell, esta tese investiga a génese dos trios de piano jazz e examina três dos mais influeciais pianostas de jazz e lideres dos mais legendários trios de piano jazz modernos: Bud Powell, Bill Evans e Keith Jarrett. Esta tese também abordará o paradoxo inerente a um sistema democrático - a expressão própria do individuo, em justaposição com a responsabilidade para com o todo – e a sua inequivoca analogia com o gestalt do trio de piano de jazz moderno. Desde a primeira gravação de um trio de jazz com pianista como líder em 1935, o trio de jazz moderno, evoluiu tornando-se um exemplo de democracia – um contexto de igualdade em que as funções rítmicas, harmónicas, e melódicas estão igualmente distribuídas entre os três instrumentistas, que são ao mesmo tempo solistas e acompanhadores. Esta tese sublinha a eficácia do trio de jazz moderno – o seu início, e porque subsiste – baseado na sua força e beleza estética; ABSTRACT: THE TWENTIETH CENTURY JAZZ PIANO TRIO – The Rise of an Iconic Jazz Paradigm by Susan Muscarella Designating Bud Powell’s 1947 recording, The Bud Powell Trio, as the modern jazz piano trio benchmark, here, this thesis traces the emergence and evolution of the pianistled, piano-bass-drums-comprised modern jazz piano trio. Beginning with a general overview of pre-Powell jazz piano styles and techniques, this thesis investigates the earliest, most salient pre-Powell jazz piano trios, and examines three seminal modern jazz pianists and leaders of legendary modern jazz piano trios, Bud Powell, Bill Evans and Keith Jarrett. This thesis also brings to the fore the paradox inherent in a democratic system – individual self-expression juxtaposed with responsibility to the whole – and its unequivocal analogy to the modern jazz piano trio gestalt. From the earliest recording of a primarily piano-dominated piano-bass-drums-comprised jazz piano trio in 1935, the modern jazz piano trio has evolved to become a paragon of democracy – an egalitarian playing field in which rhythmic, harmonic and melodic roles are evenly distributed among all three instrumentalists who have come to serve as both soloists and accompanists.