Machine Learning (ML) uses statistics to find patterns in high dimensional data. The Swedish Board of Student Finance (CSN) wants to improve the way they classify new loan takers. Using Machine Learning (ML) on data from previous loan takers can establish patterns to use on new loan takers. The aim of this study is to investigate if CSN can improve the way they classify loan takers by their ability to pay back their loan. In this study, different ML models are applied to a data set from CSN, their performance are compared and investigated by the most related factors affecting an individuals repayment rate. A data set of a total of 2032095 individuals were analysed and used in the different models. Using Random Forest (RF) for binary classification produced the best result with a sensitivity of 0.9695 and a specificity of 0.8058.

The fifth generation of mobile internet is upon us, but still, there is work to do before the new technology is fully utilized. The new generation of cellular network promises frequencies ranging from the sub – 6 GHz to 39 GHz, the latter being in the mmWave spectrum. At these frequencies, we can utilize geometrical optics to calculate the radio wave propagations. The purpose of this work is to explore how raytracing can be used to predict wireless radio wave channels and pathloss in indoor and urban environment settings. The model presented in this work explores line of sight, reflection, refraction, diffraction and scattering. The model utilizes Frii’s pathloss model for pathloss in the line-of-sight case. For reflection and refraction, Snell’s laws of reflection and refraction were used. For diffraction, the uniform theory of diffraction was used, and the scattering explored in this work was created using a physics based bidirectional reflectance diffusion function. With this we create a basic raytracing program for simple environments with potential for expansion in future work. The simple environment is a cube made in an STL file. The algorithms for the different parameter were a hybrid method of shoot- and bounce and image method for reflection, a double counting method for refraction. The transitions between different shadow boundaries were smooth. While scattering was explored, more work needs to be done to implement scattering into the code. / Den femte generationen mobilt internet är över oss, men det återstår fortfarande arbete innan den nya tekniken utnyttjas fullt ut. Den nya generationen av mobilnät lovar frekvenser som sträcker sig från sub-6 GHz till 39 GHz, där det senare ligger i millimeter-spektrumet. Vid dessa frekvenser kan vi använda geometrisk optik för att beräkna radiovågsutbredningen. Syftet med detta arbete är att utforska hur raytracing kan användas för att förutsäga trådlösa radiovågskanaler och vägförluster i inomhus- och stadsmiljö. Modellen som presenteras i detta arbete utforskar synlinje, reflektion, brytning, diffraktion och spridning. Modellen använder Friis pathloss-modell för pathloss i siktlinjefallet. För reflektion och brytning användes Snells lagar för reflektion och brytning. För diffraktion användes den enhetliga teorin om diffraktion, och spridningen som utforskades i detta arbete skapades med hjälp av en fysikbaserad dubbelriktad reflektansdiffusionsfunktion. Med detta skapar vi ett grundläggande strålspårningsprogram för enkla miljöer med potential för expansion i framtida arbete. Algoritmerna för de olika parametrarna var en hybridmetod för skjut- och studs och bildmetod för reflektion, en dubbelräkningsmetod för refraktion. Övergångarna mellan olika skugggränser var kontinuerliga. Även om spridning undersöktes, så behövs mer arbete för att implementera spidning i koden.

The use of Stereotactic Radiation Therapy (SRT) employing one large fraction of radiation, as in stereotactic radiosurgery (SRS), or few fractions of high doses, has continuously increased due to the technical development and the progress in dose delivery complemented by the positive clinical experience. The success of stereotactic radiation therapy depends on many clinical, dosimetric and radiobiological factors. For SRS in particular, the delivery of a highly conformal dose distribution to the target in one fraction allowing at the same time the sparing of the normal tissue and the critical structures is part of the basic concept of the technique. Provided that the highly accurate radiosurgical equipment available today is used, planning and delivering the prescribed dose distribution is an achievable goal, and therefore the main issue to be solved is the definition of the target. As the target volume in radiosurgery is usually defined without margins, the success of the stereotactic approach critically depends on the accurate delineation of the target which could be identified as a factor of key importance. In addition, the delineation of the Organs At Risk (OAR) is also critical. The purpose of this work was to evaluate the current degree of variability for target and OAR contouring and to establish methods for analysing multi-observer data regarding structure delineation variability. A multi-center target and OAR delineation study was initiated. Two complex and six common cases to be treated with SRS were selected and subsequently distributed to centers around the world performing Gamma Knife® radiosurgery for delineation and treatment planning. The resulting treatment plans and the corresponding delineated structures were collected and analysed.    Results showed a very high variability in contouring for four complex radiosurgery targets. Similar results indicating high variability in delineating the OAR and reporting the doses delivered to them were also reported. For the common radiosurgery targets however, a higher agreement in the delineation was observed, although lower than expected. The assessment of the quality of treatment planning for radiosurgery is usually performed with respect to the coverage of the target, the planning specificity, and dose to the sensitive structures and organs close to the target. However, physical dose conformity to the target does not guarantee the success of the treatment. The assessment of the plan quality should also be performed with respect to the clinical outcome expressed as probability of controlling the target that should be irradiated. In this respect, this study also aimed to create the framework for assessing the impact of the inaccuracy in delineating the target on the predicted treatment outcome for radiosurgery targets known for their high potential to invade the neighbouring normal tissue, using radiobiological models. In addition, radiobiological models have also been used to determine the tumour control probability accounting for the oxygenation for stereotactic radiation therapy targets. The results suggest that radiobiological modelling has the potential to add to the current knowledge in SRS by theoretically assessing the key factors that might influence the treatment outcome.

The beneficiary effects of WS2 inorganic fullerene-like nanoparticles in the lubrication industry were shown in recent years. The incorporation of the nanoparticles into lubricants (oils, greases) is however not straightforward. When two surfaces are sliding against each other and a lubricant is used, a thin layer (tribofilm) is formed on the contact area, which effects the friction process. Lubricants usually contain several additives. These additives can impair the friction reducing behaviour of the WS2 inorganic fullerene-like nanoparticles. This thesis investigates the effects of several additives in the lubrication process by analysing the tribofilms formed on the worn surfaces using energy-dispersive X-ray spectroscopy in a scanning electron microscope.

Measurements play a central role in quantum information. This thesis looksat two types: contextual measurements and symmetric measurements. Contextualityoriginates from the Kochen-Specker theorem about hidden variablemodels and has recently undergone a subtle shift in its manifestation. Symmetricmeasurements are characterised by the regular polytopes they formin Bloch space (the vector space containing all density matrices) and are thesubject of several investigations into their existence in all dimensions.We often describe measurements by the vectors in Hilbert space ontowhich our operators project. In this sense, both contextual and symmetricmeasurements are connected to special sets of vectors. These vectors areoften special for another reason: they form congurations in a given incidencegeometry.In this thesis, we aim to show various connections between congurationsand measurements in quantum information. The congurations discussedhere would have been well-known to 19th and 20th century geometers andwe show they are relevant for advances in quantum theory today. Specically,the Hesse and Reye congurations provide proofs of measurement contextuality,both in its original form and its newer guise. The Hesse congurationalso ties together dierent types of symmetric measurements in dimension3called SICs and MUBswhile giving insights into the group theoreticalproperties of higher dimensional symmetric measurements.

For the successful deployment of advanced nuclear systems and for optimization of current reactor designs, high quality and accurate nuclear data are required. Before nuclear data can be used in applications, they are first evaluated, benchmarked against integral experiments and then converted into formats usable for applications. The evaluation process in the past was usually done by using differential experimental data which was then complimented with nuclear model calculations. This trend is fast changing because of increase in computational power and tremendous improvements in nuclear reaction theories over the last decade. Since these model codes are not perfect, they are usually validated against a large set of experimental data. However, since these experiments are themselves not exact, the calculated quantities of model codes such as cross sections, angular distributions etc., contain uncertainties. A major source of uncertainty being the input parameters to these model codes. Since nuclear data are used in reactor transport codes asinput for simulations, the output of transport codes ultimately contain uncertainties due to these data. Quantifying these uncertainties is therefore important for reactor safety assessment and also for deciding where additional efforts could be taken to reduce further, these uncertainties. Until recently, these uncertainties were mostly propagated using the generalized perturbation theory. With the increase in computational power however, more exact methods based on Monte Carlo are now possible. In the Nuclear Research and Consultancy Group (NRG), Petten, the Netherlands, a new method called ’Total Monte carlo (TMC)’ has been developed for nuclear data evaluation and uncertainty propagation. An advantage of this approach is that, it eliminates the use of covariances and the assumption of linearity that is used in the perturbation approach. In this work, we have applied the TMC methodology for assessing the impact of nuclear data uncertainties on reactor macroscopic parameters of the European Lead Cooled Training Reactor (ELECTRA). ELECTRA has been proposed within the GEN-IV initiative within Sweden. As part of the work, the uncertainties of plutonium isotopes and americium within the fuel, uncertainties of the lead isotopes within the coolant and some structural materials of importance have been investigated at the beginning of life. For the actinides, large uncertainties were observed in the k-eff due to Pu-238, 239, 240 nuclear data while for the lead coolant, the uncertainty in the k-eff for all the lead isotopes except for Pb-204 were large with significant contribution coming from Pb-208. The dominant contributions to the uncertainty in the k-eff came from uncertainties in the resonance parameters for Pb-208. Also, before the final product of an evaluation is released, evaluated data are tested against a large set of integral benchmark experiments. Since these benchmarks differ in geometry, type, material composition and neutron spectrum, their selection for specific applications is normally tedious and not straight forward. As a further objective in this thesis, methodologies for benchmark selection based the TMC method have been developed. This method has also been applied for nuclear data uncertainty reduction using integral benchmarks. From the results obtained, it was observed that by including criticality benchmark experiment information using a binary accept/reject method, a 40% and 20% reduction in nuclear data uncertainty in the k-eff was achieved for Pu-239 and Pu-240 respectively for ELECTRA.

The usage of neural networks to derive physical laws without any kind of pre-existing bias is a promising but relatively new field with the long-term goal to construct an artificial intelligence physicist that is able to derive physical laws from experimental data. In this project, a step is taken in the direction of solving complex problems by tackling the double pendulum, the simplest chaotic system. To do so, a neural network architecture, adapted from previous work, is used to find the relevant parameters of the system in multiple configurations of the pendulum. Afterwards, the possibility of a neural network derived general solution of the problem is discussed through the relevant aspects that increase its complexity.

Neutron detectors in the form of liquid organic scintillators are widely used for nuclear disarmament purposes. The main purpose of this thesis is to characterize the neutron detection efficiency of the organic liquid scintillator EJ-309 using Monte Carlo based simulations in the simulation program Geant4. The Monte Carlo simulation code was benchmarked using recorded light output data from a laboratory experiment and it was confirmed that the simulation code could reproduce data from this experiment. In order to characterize the efficiency of the neutron detector, the benchmarked code was used to simulate light output distributions of 241-Am-Be and Cf-252, an (α, n) and a spontaneous fission source respectively, where it was found that the two different types of sources could be distinguished. However, distinguishment between two different spontaneous fission sources was not possible, as simulations of the light output distribution of another spontaneous fission source, Pu-240, was similar to the one for Cf-252. The light output distributions from the simulations were compared to earlier research where good agreements were obtained. A quantitative study was also performed to determine how many detected neutrons were needed to properly be able to distinguish between the two sources 241-Am-Be and Cf-252, and the number was found to be 560. This value was considered low relative to the number of detected neutrons in the simulations, and hence, more quantitative studies are needed in order to verify this result. The absolute and intrinsic detection efficiency were also determined for varying threshold values of neutron energy.

It could be argued that a renaissance is taking place for the global interest in nuclear power - especially for the development of the next generation of reactor systems that aim at solving the problems with contemporary energy production with nuclear technology. One such proposed concept for Gen IV nuclear power is a Molten Salt Reactor (MSR), in which the nuclear fuel is in solution of a salt - often based on fluorine or chlorine. The historical experience of MSRs is limited in both scope and time, but significant developments have been made in the past few years. In order for such technologies to become commercially viable, it is important that investigations are conducted into the behaviour of the materials involved. In this work, the propagation of uncertainty in nuclear data for the isotope F-19 is investigated in a fluorine-based MSR. Uncertainty quantification is important in reactor physics as calculations that are based on Best Estimate Plus Uncertainty often leave more margin to the regulatory requirements compared to conservative calculations. The quantity of interest is the reactivity of the system at hand, and three different levels of moderation were investigated in order to observe any differences that the hardness of the neutron spectrum might impose on the reactivity uncertainty. In addition, the individual neutron-nucleus interaction channels are analyzed separately in order to deduce which channels that contribute the most to the reactivity uncertainty. The nuclear reactor cores were simu- lated with the Monte Carlo-based neutron transport code OpenMC, and the uncertainty quantification was performed using the Total Monte Carlo (TMC) methodology with perturbed nuclear data that was generated using the tools SANDY and NJOY. The results from the uncertainty quantification showed that the nuclear data uncertainty from F-19 gave rise to reactivity uncertainties of approximately 60-200 pcm, depending on the amount of moderation rods that were inserted. The less moderated systems were more susceptible to the F-19 uncertainties, which could be explained by the fact that those systems gave more room for the fluorine- based fuel, and hence they would be more exposed to the uncertainties of F-19. It was also observed that the elastic scattering, neutron capture, and alpha production reaction channels contributed the most to the uncertainties in the most moderated reactor, while the least moderated reactor was most susceptible to the uncertainties due to elastic scattering, inelastic scattering, and alpha production. These findings show that significant improvements needs to be made in the measurements and evaluations of nuclear data for F-19 if it is to be implemented in MSRs of the future. / Intresset för kärnkraft har under de senaste åren utvecklats enormt globalt - särskilt med avseende på utveckling av nästa generation av reaktorsystem som har målen att lösa de problem som finns med dagens energiproduktion via kärnteknik. Ett sådant föreslaget koncept för Fjärde generationens kärnkraft är saltsmältereaktorn (eng. Molten Salt Reactor, MSR), i vilket kärnbränslet är upplöst i reaktorn i ett salt som oftast är baserat på fluor eller klor. Den historiska erfarenheten av MSR:er är knapphändig, men stora framsteg har gjorts under de senaste åren. För att tekniker som denna ska kunna blir kommersiellt genomförbara måste utredningar göras för att förstå hur exempelvis det använda materialet beter sig. I detta arbete propageras osäkerheterna i kärndata för isotopen F-19 genom en fluor-baserad MSR. Os- äkerhetskvantifiering är viktigt i reaktorfysik eftersom beräkningar som är baserade på Bästa Uppskat- tning Plus Osäkerhet (eng. Best Estimate Plus Uncertainty, BEPU) oftast lämnar en större marginal till regulatoriska begräsningar än vad konservativa beräkningar gör. Storheten av intresse i detta ar- bete är reaktiviteten av reaktorhärden, och tre olika nivåer på moderering användes för att kunna observera skillnader i reaktivitetsosäkerheter som orsakas av neutronspektrumets hårdhet. Dessutom undersöktes de individuella reaktionskanalerna för neutron-kärna-interaktioner för att kunna dra en slutsats kring vilka reaktionskanaler som bidrar mest till reaktivitetsosäkerheterna. Reaktorhärdarna simulerades med den Monte Carlo-baserade neutrontransportkoden OpenMC, och osäkerhetskvanti- fieringen gjordes genom Total Monte Carlo (TMC) metodologin med kärndata som genererades med verktygen SANDY och NJOY. Resultaten från osäkerhetskvantifieringen visade att kärndataosäkerheterna från F-19 gav upphov till reaktivitetsosäkerheter på ungefär 60-200 pcm, beroende på vilken modereringsgrad som användes. De mindre modererade systemen var mer mottagliga till osäkerheten hos F-19, vilket skulle kunna förklaras genom att dessa system gav mer plats för det fluor-baserade bränslet, vilket därmed leder till att de är exponerade till osäkerheterna för F-19 i en högre grad. Det observerades också att reaktionskanalerna elastisk spridning, neutroninfång samt alfa-produktion hade stört inverkan på reaktivitetsosäkerheterna för de mest modererade systemen, medan de minst modererade systemen var mer mottagliga till osäkerheterna som orsakades av elastisk spridning, inelastisk spridning och alfa-produktion. Dessa resultat visar att förbättringar måste göras i mätningarna samt evalueringarna av kärndata för F-19 om ämnet ska implementeras i MSR:er i framtiden.

In the Born-Oppenheimer approximation for a quantum system the emergent synthetic magnetic field can be seen as generated by monopoles at points of degeneracy, in full analogue to the synthetic fields generating the geometric phase of adiabatically evolving quantum systems. The plausibility of using these synthetic magnetic monopoles as a means to study magnetic monopole dynamics in the absence of fundamental magnetic monopoles has been explored. A bipartite spin system consisting of a dumbbell translating and rotating through space has been modelled, and full equations of motion in the presence of an external magnetic field have been derived. A collection of scripts for numerical evaluation of these equations of motion were subsequently developed, and further put to use in sample simulations for a small range of parameters. The results demonstrate non-negligible perturbations to the centre of mass motion when compared to motion not considering the Born-Oppenheimer synthetic fields, for dumbbell masses of small but not unrealistic proportions. The problems inherent in this approach to elucidating motion in magnetic monopole fields are discussed, but the method should not yet be dismissed until further investigations have been made. / Under Born-Oppenheimer-approximationen för ett kvantsystem kan det emergenta syntetiska magnetfältet ses som alstrat av monopoler vid degenerationspunkter, helt analogt med de syntetiska fält som genererar den geometriska fasen vid adiabatisk utveckling av kvantsystem. Möjligheten att använda dessa syntetiska magnetiska monopoler för att studera dynamiken från verkan av en magnetisk monopol, trots att fundamentala magnetiska monopoler ej observerats, har utforskats. Ett tvådelat spinsystem beståendes av en hantel som translaterar och roterar genom rummet har modellerats, och fullständiga rörelseekvationer i närvaron av ett yttre magnetfält har härletts. Kod till ändamålet att utvärdera dessa rörelseekvationer har därpå utvecklats, och vidare nyttjats för att simulera rörelsen för ett stickprov av parametrar. Resultaten visar på ej försumbara perturbationer av masscentrums rörelse vid jämförelse med rörelse utan hänsyn till de syntetiska Born-Oppenheimer-fälten, för hantlar av liten men inte orealistiskt liten massa. Problemen och komplikationerna för det här angreppssättet till att utforska rörelse genom magnetiska monopolers fält diskuteras, men metoden bör ej ännu avvisas innan vidare undersökning har genomförts.