Durant les dernières décennies, l’évolution de la puissance de calcul a conduit au développement de codes de simulation en physique des réacteurs de plus en plus prédictifs pour la modélisation du comportement d’un réacteur nucléaire en situation de fonctionnement normal et accidentel. Un cadre d’analyse d’incertitudes cohérent avec l’utilisation de modélisations Best Estimate (BE) a été développé. On parle d’approche Best Estimate Plus Uncertain-ties (BEPU) et cette approche donne lieu `a de nombreux travaux de R&D à l’international en simulation numérique. Dans cette thèse, on étudie la quantification d’incertitudes multi-physiques dans le cas d’un transitoire d’ éjection de Grappe de contrôle (REA- Rod Ejection Accident) dans un Réacteur à Eau Pressurisée (REP). La modélisation BE actuellement disponible au CEA est réalisée en couplant les codes APOLLO3 R (netronique) et FLICA4 (thermohydraulique-thermique du combustible) dans l’environnement SALOME/CORPUS. Dans la première partie de la thèse, on examine différents outils statistiques disponibles dans la littérature scientifique dont la réduction de dimension, l’analyse de sensibilité globale, des modèles de substitution et la construction de plans d’expérience. On utilise ces outils pour développer une méthodologie de quantification d’incertitudes. Dans la deuxième partie de la thèse, on améliore la modélisation du comportement du combustible. Un couplage Best Effort pour la simulation d’un transitoire REA est disponible au CEA. Il comprend le code ALCYONE V1.4 qui permet une modélisation fine du comportement thermomécanique du combustible. Cependant, l’utilisation d’une telle modélisation conduit à une augmentation significative du temps de calcul ce qui rend actuellement difficile la réalisation d’une analyse d’incertitudes. Pour cela, une méthodologie de calibrage d’un modèle analytique simplifié pour le transfert de chaleur pastille-gaine basé sur des calculs ALCYONE V1.4 découplés a été développée. Le modèle calibré est finalement intégré dans la modélisation BE pour améliorer sa prédictivité. Ces deux méthodologies sont maquettées initialement sur un cœur de petite échelle représentatif d’un REP puis appliquées sur un cœur REP à l’échelle 1 dans le cadre d’une analyse multi-physique d’un transitoire REA. / The computational advancements of the last decades lead to the development of numerical codes for simulating the reactor physics with increa-sing predictivity allowing the modeling of the beha-vior of a nuclear reactor under both normal and acci-dental conditions. An uncertainty analysis framework consistent with Best Estimate (BE) codes was develo-ped in order to take into account the different sources of uncertainties. This framework is called Best Esti-mate Plus Uncertainties (BEPU) and is currently a field of increasing research internationally. In this the-sis we study the multi-physics uncertainty quantifi-cation for Rod Ejection Accident (REA) in Pressuri-zed Water Reactors (PWR). The BE modeling avai-lable in CEA is used with a coupling of APOLLO3 (neutronics) and FLICA4 (thermal-hydraulics and fuel-thermal) in the framework of SALOME/CORPUS tool. In the first part of the thesis, we explore different statistical tools available in the scientific literature including: dimension reduction, global sensitivity analy-sis, surrogate modeling and design of experiments. We then use them in order to develop an uncer-tainty quantification methodology. In the second part of the thesis, we improve the BE modeling in terms of its uncertainty representation. A Best Effort coupling scheme for REA analysis is available at CEA. This in-cludes ALCYONE V1.4 code for a detailed modeling of fuel-thermomechanics behavior. However, the use of such modeling increases significantly the compu-tational cost for a REA transient rendering the uncer-tainty analysis prohibited. To this purpose, we deve-lop a methodology for calibrating a simplified analytic gap heat transfer model using decoupled ALCYONE V1.4 REA calculations. The calibrated model is finally used to improve the previous BE modeling. Both de-veloped methodologies are tested initially on a small scale core representative of a PWR and then applied on a large scale PWR core.

This master`s thesis deals with the performance measurement of Stredoslovenská energetika, a.s. The first section explains what performance is and describes the theoretical basis, which relates with the methods of measurement and analysis of the company’s current state. In the following part there is the profile of the company with its current situation. The evaluation of the performance SSE, a.s. is measured by START PLUS model. This model identifies the strengths and the areas which need improvement and offers suggestions for improving the company’s performance.

Klinisk kemiska analyser har hög klinisk relevans. I serum/plasma kan olika parametrar kvantifieras. Dessa parametrar kan vara proteiner, enzymer, joner, metaller, lipider och kolhydrater. Med hjälp av referensintervall kan veterinärer ställa diagnos, följa behandling och sjukdomsförlopp. Parametrar detekteras med olika analysprinciper/metoder; kolorimetri, immunturbidimetri, enzymatisk metod och potentiometri. Djursjukhuset, AniCura, i Hässleholm mottagas både hundar och katter. Cirka 120 kemianalyser analyseras varje dag. AniCura har köpt in ett nytt våtkemiiinstrument, Indiko Plus, som ska ersätta Cobas C111. Med Indiko Plus tillkommer fler provpositioner, 8 nya analyser, ökad kapaciteten och underlättad användning.  Syftet med denna studie var att jämföra instrumenten och ta fram eget referensintervall som jämfördes med referensintervall framtaget av Thermo Fisher. Verifiering av det nya instrumentet genomfördes med precisionsstudie och linjäritetsstudie. Provtagning på hundar och katter utfördes av personal på AniCura. Jämförelsen gjordes med patientprover och referensintervall togs fram med hjälp av prover från friska hundar och katter. Jämförelsen visade att 9 av 13 analyser hade statistisk signifikant skillnad. Orsaken till det beror troligen på skillnaden av reagens, instrumentens ålder och tid mellan mätningar. Ett nytt referensintervall utarbetades och skiljde sig inte mycket från Thermo Fishers intervall. Vidare validering på grund av liten population rekommenderades. Precisionen för Indiko Plus blev godkänd. Linjäriteten blev icke linjär och berodde troligen på en dålig pipett och bör göras om. / Clinical chemical analyzes has high clinical relevance. In serum/plasma, different parameters can be quantified. Parameters can be proteins, enzymes, ions, metals, lipids, and carbohydrates. With reference intervals, veterinarians can set diagnosis, follow treatment and development of the disease. Parameters are detected with different analysis principles/methods; colorimetry, immunoturbidimetry, enzymatic method and potentiometry. The animal hospital, AniCura, in Hässleholm accept dogs and cats. About 120 chemical analyzes are analyzed every day. AniCura purchased a new instrument, Indiko Plus, which will replace Cobas C111. Indiko Plus provide, more sample positions, 8 new analyzes, increased capacity, and facilitated use. The purpose of this study was to compare the instruments and produce a new reference interval which was compared to the reference interval provided by Thermo Fisher. To verify Indiko Plus, a precision and linearity study were conducted. Blood sampling of dogs and cats was performed by staff at AniCura. The comparison was made with patient samples and the reference intervals were obtained using samples from healthy animals. The comparison showed 9 of 13 analyzes had a statistically significant difference. The reason for this is probably due to the difference in reagents, the age of the instruments and the time between measurements. A new reference interval was developed and did not differ much from the Thermo Fisher interval. Further validation due to low population was recommended. The precision for Indiko Plus was approved. The linearity study shows not linear trend but was likely due to a bad pipette and should be redone.

The thesis focuses on the British approach to EMU fiscal rules. Theoretical and conceptual framework is based on the theory of differentiated integration dealing with a various forms of closer cooperation among EU Member States. The research is built on the theoretical approach as described by Frank Schimmelfennig, Dirk Leuffen and Berthold Rittberger. Authors provide an analysis of conditions of differentiated integration within the EU. The aim of this thesis is to answer the question of why the United Kingdom decided not to participate in the fiscal compact and, consequently, did not allow further differentiated integration within the treaties (which led to an arrangement outside the EU legal framework) while accepting the initial fiscal rules agreed in the 1990s as well as other crisis reforms of EU fiscal governance. The first part describes Maastricht and SGP fiscal rules and analyses both the UK's position within EMU and the SGP provisions applied to the UK. The next chapter examines the British approach towards both the SGP revisions and the Euro Plus Pact. The third part analyses the new fiscal compact and explains why the UK decided to use the veto. In conclusion, the thesis provides an answer to the research question as set down in the introduction.

Indiana University-Purdue University Indianapolis (IUPUI) / Purpose: This in vitro study aims to analyze the effect of combining two ADA-recommended professionally applied 1:1 Chlorhexidine/Thymol varnish ((Cervitec Plus)) and professionally prescribed 5,000 ppm fluoride toothpaste ((PreviDent 5000 Plus)) on reducing lesion depth and increasing mineral content Materials & Methods: Forty-eight dentin specimens were randomly distributed into four treatment groups (n=12 per treatment). Biofilms of Streptococcus mutans and Candida albicans were created on the polished surfaces of bovine root dentin specimens (n=12 per treatment). 1:1 Chlorhexidine/Thymol varnish was applied once then the tested 5,000 ppm fluoride toothpaste was applied for 120 seconds twice daily over the course of 2 days. Tested groups were: (1) 1:1 Chlorhexidine/Thymol varnish ((Cervitec Plus)) (C/T). (2) 5,000 ppm F toothpaste ((PreviDent 5000 Plus)) (F). (3) Combination of 1:1 Chlorhexidine/Thymol varnish ((Cervitec Plus)) & 5000 ppm F toothpaste ((PreviDent 5000 Plus)) (C/T+F). (4) Deionized water (DIW) as control group. Biofilms were analyzed for biofilm dry weight. Dentin specimens were analyzed using transversal microradiography (TMR) for mineral content change and lesion depth. PH data was analyzed using two-way ANOVA. Total biofilm dry weight data was analyzed using one-way ANOVA. Integrated mineral loss and lesion depth data was analyzed using two-way ANOVA All pair-wise comparisons from ANOVA analysis were made using Fisher’s Protected Least Significant Differences to control the overall significance level at 5%. Results: Treatment with (C/T+F) resulted in higher mean pH values compared to the control group (DIW) and (F) group. The average pH values of group (C/T) were not statistically different than group (C/T+F). the biomass of the combined S. mutans & C. albicans biofilm among all the groups were not significantly different. (DIW) presented significantly deeper lesions for both surfaces (sound &demineralized) when compared to (F) (P=0.0118), (C/T) (P=0.0002), and (C/T+F) (P<.0001). The sound surfaces for the specimens for group (C/T) and Group (F) showed superficial lesion depth. However, the sound surfaces of specimens treated with (C/T+F) showed the most superficial depth. Due to mineral gain, the demineralized surfaces of the specimens of both (C/T) & (C/T+F) showed a decrease in the lesion depth. Conclusion: Within the limitations of our study. The combination of 5,000 ppm fluoride toothpaste and CHX/Thymol had no significant effect on mineral content. However, the combination had a considerable effect on lesion depth reduction.

El presente trabajo de investigación académica está basado en el análisis de las acciones de la promoción digital en relación con el proceso de compra de ropa Plus Size en mujeres de 18 a 34 años de NSE B de Lima Metropolitana. En primer lugar, se dan a conocer cuales son las estrategias SEO y SEM, así mismo detalla cuales son las acciones comprendidas dentro de estas estrategias. Además, se investiga sobre la reacción del target a estas acciones y estímulos. En segundo lugar, se tomó muy en cuenta el análisis del proceso de compra de las mujeres del segmento, descubriéndose que el target prefiere realizar compras por internet antes que las tiendas físicas. Finalmente, se contrastaron ambas variables y se evaluó la influencia de las diferentes estrategias de marketing digital en las etapas del proceso de compra del segmento Plus Size. Adicionalmente, se presenta el contraste de los resultados con la hipótesis planteada y conclusiones de la investigación. / This academic research work is based on the analysis of the actions of digital promotion in relation to the process of buying Plus Size clothing in women from 18 to 34 years of age from NSE B of Metropolitan Lima. First of all, they are disclosed which are the SEO and SEM strategies, as well as detailing what are the actions included within these strategies. In addition, the target's reaction to these actions and stimuli is investigated. Secondly, I am very aware of the analysis of the purchasing process of women in the segment, discovering that the objective of making purchases online before physical stores. Finally, both variables were compared and the influence of the different digital marketing strategies in the stages of the Plus Size segment purchasing process was evaluated. In addition, the contrast of the results with the hypotheses and the conclusions of the investigation is presented. / Trabajo de investigación

Les problèmes de contrôle stochastique optimal à horizon fini forment une classe de problèmes de contrôle optimal où interviennent des processus stochastiques considérés sur un intervalle de temps borné. Tout comme beaucoup de problème de contrôle optimal, ces problèmes sont résolus en utilisant le principe de la programmation dynamique qui induit une équation aux dérivées partielles (EDP) appelée équation d'Hamilton-Jacobi-Bellman. Les méthodes basées sur la discrétisation de l’espace sous forme de grille, les méthodes probabilistes ou plus récemment les méthodes max-plus peuvent alors être utilisées pour résoudre cette équation. Cependant, le premier type de méthode est mis en défaut quand un espace à dimension grande est considéré à cause de la malédiction de la dimension tandis que le deuxième type de méthode ne permettait jusqu'ici que de résoudre des problèmes où la non linéarité de l'équation aux dérivées partielles par rapport à la Hessienne n'est pas trop forte. Quant au troisième type de méthode, il entraine une explosion de la complexité de la fonction valeur. Nous introduisons dans cette thèse deux nouveaux schémas probabilistes permettant d'agrandir la classe des problèmes pouvant être résolus par les méthodes probabilistes. L'une est adaptée aux EDP à coefficients bornés tandis que l'autre peut être appliqué aux EDP à coefficients bornés ou non bornés. Nous prouvons la convergence des deux schémas probabilistes et obtenons des estimées de l'erreur de convergence dans le cas d'EDP à coefficients bornés. Nous donnons également quelques résultats sur le comportement du deuxième schéma dans le cas d'EDP à coefficients non bornés. Ensuite, nous introduisons une méthode complètement nouvelle pour résoudre les problèmes de contrôle stochastique optimal à horizon fini que nous appelons la méthode max-plus probabiliste. Elle permet d'utiliser le caractère non linéaire des méthodes max-plus dans un contexte probabiliste tout en contrôlant la complexité de la fonction valeur. Une application au calcul du prix de sur-réplication d'une option dans un modèle de corrélation incertaine est donnée dans le cas d’un espace à dimension 2 et 5. / Stochastic optimal control problems with finite horizon are a class of optimal control problems where intervene stochastic processes in a bounded time. As many optimal control problems, they are often solved using a dynamic programming approach which results in a second order Partial Differential Equation (PDE) called the Hamilton-Jacobi-Bellman equation. Grid-based methods, probabilistic methods or more recently max-plus methods can be used then to solve this PDE. However, the first type of methods default in a space of high dimension because of the curse of dimensionality while the second type of methods allowed till now to solve only problems where the nonlinearity of the PDE with respect to the second order derivatives is not very high. As for the third type of method, it results in an explosion of the complexity of the value function. We introduce two new probabilistic schemes in order to enlarge the class of problems that can be solved with probabilistic methods. One is adapted to PDE with bounded coefficients while the other can be applied to PDE with bounded or unbounded coefficients. We prove the convergence of the two probabilistic scheme and obtain error estimates in the case of a PDE with bounded coefficients. We also give some results about the behavior of the second probabilistic scheme in the case of a PDE with unbounded coefficients. After that, we introduce a completely new type of method to solve stochastic optimal control problems with finite horizon that we call the max-plus probabilistic method. It allows to add the non linearity feature of max-plus methods to a probabilistic method while controlling the complexity of the value function. An application to the computation of the optimal super replication price of an option in an uncertain correlation model is given in a 5 dimensional space.

Max-plus代数は，実数全体に無限小元を付加した集合に，加法として最大値をとる演算，乗法として通常の加法を考えた代数系である．本論文では，max-plus線形方程式に対するCramerの公式の類似物を用いて，線形方程式の解空間の基底が構成できることを示した．さらに固有値問題に関連して，max-plus行列の固有ベクトルの概念を2通りの観点から拡張した． / The max-plus algebra is the semiring with addition "max" and multiplication "+". In the present thesis, the author gives a combinatorial characterization of solutions of linear systems in terms of the max-plus Cramer's rule. Further, the author extends the concept of eigenvectors of max-plus matrices from two different perspectives. / 博士(理学) / Doctor of Philosophy in Science / 同志社大学 / Doshisha University

Streaming services have had a significant gain in popularity over the last decade, with an increasing number of subscribers. Along with digitalization, more households seek accessible entertainment that allows them to choose the type of content they wish to watch. However, as streaming services have grown in popularity, the content library of those services has been increasing, leaving users overwhelmed when deciding what to watch next. A recommendation system is used by the majority of streaming providers to assist its customers find appropriate material to watch. This recommendation system suggests suitable material based on the user's collected data. However, the recommendation systems suggestions are not always perceived to be relevant. The purpose of this thesis is to identify potential perceived problems with the recommendation systems and their causes. This was accomplished by investigating how consumers of those services experience them by conducting a survey and interviews. The identified problems have then been compiled into a basis. The goal of this thesis is to provide a basis that may benefit streaming services, their consumers and be of use by academics to conduct further research within the field. The basis indicates that there are several perceived problems with the current recommendation systems. A total of nine problems have been documented and discussed. / Streamingtjänster har haft en betydande ökning i popularitet under det senaste årtiondet, med ett ökande antal användare. Tillsammans med digitaliseringen söker fler hushåll tillgänglig underhållning som gör att de kan välja vilken typ av innehåll de vill titta på. Men eftersom streamingtjänster har ökat i popularitet, har utbudet för dessa tjänster ökat, vilket gör användarna överväldigade när de bestämmer sig för vad de ska titta på härnäst. Rekommendationssystemen används av majoriteten av streamingtjänsterna för att hjälpa sina kunder att hitta lämpligt material att titta på. Rekommendations systemen föreslår lämpligt material baserat på användarens beteende mönster. Rekommendations systemets förslag uppfattas dock inte alltid vara relevanta. Syftet med denna uppsats är att identifiera potentiella problem som användare uppfattar av rekommendations systemen och vad som kan orsaka dessa problem. Detta åstadkoms genom att undersöka hur konsumenter av dessa tjänster upplever dem genom att genomföra en undersökning samt ett antal intervjuer. De identifierade problemen har sedan sammanställts till ett underlag. Målet med detta examensarbete är att ge ett underlag som kan gynna streamingtjänster, deras konsumenter och vara till nytta för akademiker för att genomföra ytterligare studier inom området. Underlaget tyder på att det finns flera problem med nuvarande rekommendationssystem som upplevs av användare. Totalt har nio problem dokumenterats och diskuterats.

I den här rapporten utvärderas det nya konceptet Power &amp; Biomass to Liquid (PBtL). PBtL är ett alternativ till den tidigare och mer etablerade Biomass to Liquid (BtL) processen. Med PBtL förbättras utbytet av kol jämfört med BtL genom att elektricitet läggs till i processen. Elektriciteten används för att producera H2, som används för att höja H2/CO förhållandet istället för att använda WGS som i vanlig BtL process. Rapporten är en del i ett större PBtL projekt som bedrivits vid Institutt for kjemisk prosessteknologi vid NTNU och på SINTEF. Utvärderingen utfördes genom flera simuleringar av lågtemperaturs Fischer-Tropsch reaktorer i simuleringsprogrammet Aspen Plus. Omvandling och katalytiska reaktorer utvecklades och togs fram i programmet. Produktfördelningen i omvandlingsreaktorn modellerades med ASF distribution theory tillsammans med en metod för sammanslagning av högre kolväten. Fördelningen av paraffiner, olefiner och oxygenater baserades på experimentella resultat från Shafer et al. som studerade en slurryreaktor under liknande förhållanden. Den kinetiska reaktorn modellerades med en variant av ASF fördelningsteori kallad ”consorted vinylene mechanism” från Rytter och Holmen. Reaktorerna adderades till förgasningsprocess, som utvecklats tidigare av PBtL gruppen, I förgasningsprocessen förgasas biomassa till syntesgas, dvs H2 och CO. För att möjliggöra en utvärdering av det efterföljande steget med separering av vax, mellandestillat och lättare kolväten så antogs en väl fungerande separation av Fischer-Tropsch produkterna. En enklare separation av med flash förångning gjordes också, dels för fortsättningen av PBtL processen och för att kunna studera tailgasrecirkulering. Ett mindre bidrag var studier av en torkningsprocess för biomassa innan inloppet till förgasningsprocessen. PBtL konceptet diskuterades även ur ett praktiskt perspektiv.  Resultaten visar att vid driftbetingelser på 210 °C, 25 bar och H2/CO = 1,95 så gav omvandlingsreaktorn en kolselektivitet för CH4 respektive C5+ på 14,77 respektive 75,40 mol% C. Högre temperatur, tryck och H2/CO förhållande i reaktorn resulterar i en högre kolselektivitet mot lägre kolväten. Vid samma driftbetingelser gav den katalysreaktorn en kolselektivitet för CH4 respektive C5+ på 7,612 respektive 86,00 mol% C. Resultaten visar att C8-C16 produktionen var högre än C17+ med avseende på molflöde men lägre beträffande massflöde för katalysreaktorn. Generellt så ökar kolselektiviteten med ökande kolnummer till ett maximum runt 13 för att sedan minska. / In this report, the new Power &amp; Biomass to Liquid (PBtL) concept was evaluated. The PBtL concept is a new alternative to the more well-established Biomass to Liquid (BtL) concept where electricity is added to the process. The main purpose for developing the PBtL is that the BtL process exhibits poor carbon efficiency compared to the PBtL process. The electricity here is used to produce H2 in electrolysis. The report is part of a larger PBtL project pursued for several years at the Department of Chemical Engineering at NTNU and SINTEF. The evaluation was done by simulating different types of low temperature Fischer-Tropsch reactors in simulation software Aspen Plus. A conversion reactor and a kinetic reactor was developed. A conversion reactor based on the result from the kinetic reactor was also developed.  The conversion-based reactor was modeled with the ASF distribution theory which describes the distribution of products formed in Fischer-Tropsch synthesis along with a method of lumping higher hydrocarbons. The distribution between paraffins, olefins and oxygenates was based on experimental data from Shafer et al. with similar operating condition with a Slurry reactor. The kinetic-based reactor was modeled with ASF distribution theory with a consorted vinylene mechanism previously described in Rytter and Holmen. The reactors were added to a process for which the biomass gasification section had previously been developed by the PBtL group. The Fischer-Tropsch products were as well separated in order to evaluate the subsequent step of separation of waxes, middle distillate and lighter hydrocarbons. This enabled the option of recycling of tail gas to the Fischer-Tropsch reactor to be evaluated. A smaller contribution included addition of a biomass dryer prior the biomass gasification section. The PBtL concept is also shortly discussed from a practical point-of-view.  It was found that for the operating condition of 210 °C, 25 bar and H2/CO = 1.95 for the conversion-based reactor yielded a carbon selectivity towards CH4 and C5+ of 14.77 and 75.40 mol C% respectively. For the same operating condition, the kinetic-based reactor yield a carbon selectivity towards CH4 and C5+ of 7.612 and 86.00 mol C% respectively. It could be seen from the conversion-based reactor that elevating temperature, pressure and H2/CO (to a certain extent) results in higher carbon selectivity towards lower hydrocarbons. From the product separation with the kinetic reactor, it was observed that C8-C16 production was higher than the C17+ production in terms of mole flow but lower in terms of mass flow. For both models, carbon selectivity increases with carbon number and peaks around carbon number 13 and then starts to decrease.