Breast Cancer is the second most common cancer in Sweden. More treatment options are available with a higher chance of survival if Breast cancer is diagnosed early. WHO has recommended performing Breast Cancer screenings on women even before they develop any symptoms. In Sweden, all women between the age of 40 and 74 years are called for a Breast Cancer Screening examination every two years. However, all of the called out women are not appearing for Breast screening therefore the purpose of this degree project is to analyze the possible benefits of setting up Mobile Mammography Units (MMUs) in Sweden. This thesis focuses on need-analysis, cost-analysis, image quality, and quality of care in MMUs compared to fixed units and if MMUs can be used as a tool for dispersal of knowledge regarding preventive care for cancer. Further, long waiting queues is one of the major concern and underperformed area in Sweden. Since the enactment of Cancer Care Pathways in Sweden, long waiting queues for patients having non-cancer diseases have been the most frequently mentioned risk in the regional status reports. This study, therefore, aims to analyze if mammography screening in hospitals is adding up to the long waiting ques for other patients or not. This study followed a qualitative setting where several semi-structured interviews were conducted with members of the National workgroup of mammography in Sweden, and an abductive inductive approach was followed for the data collection and analysis. The results of the study concluded that Mobile mammography units are only beneficial for areas with demographic issues. Mobile mammography units are resource-demanding therefore it is better to invest in fixed units for areas where there are no traveling issues. The mammography unit in a hospital is a separate department and does not affect any other patient by being in the hospital hence not resulting in the “Crowding out effect”. Regarding the Image quality and quality of care, it is similar in both the fixed and mobile units, apart from the quality of care for disabled women who must visit the fixed unit and hence are deprived of close care. A proper cost-analysis, with exact figures, for both mobile and fixed mammography units was not found during this study therefore it could not be concluded if MMUs are cost-effective or not. Also, if Mobile Mammography Units are to be used as a tool for the dispersal of knowledge regarding the prevention of cancer, there will be a need to hire extra staff who can perform this job. / Bröstcancer är den näst vanligaste cancerformen i Sverige. Fler behandlingsalternativ finns tillgängliga med högre chans att överleva om bröstcancer diagnostiseras tidigt. WHO har rekommenderat att utföra bröstcancerscreeningar på kvinnor även innan de utvecklar några symtom. I Sverige kallas alla kvinnor mellan 40 och 74 år till en bröstcancerundersökning vartannat år. Alla kvinnor som bjuds in till undersökning dyker dock inte upp, därför är syftet med detta examensarbete att analysera möjliga fördelar med att sätta upp mobila mammografienheter (MMU) i Sverige. Långa väntetider är ett stort problem i Sverige, även för dessa patienter som inte har cancer eller tecken på cancer. Denna studie syftar därför till att analysera om mammografiscreening på sjukhus ökar eller minskar de långa väntetiderna för andra patienter som inte är drabbade av cancern. Resultat av studien visar att mobila mammografienheter endast är fördelaktiga för områden med demografiska problem. Mobila mammografienheter är resurskrävande, därför är det bättre att investera i fasta enheter för områden där det inte finns några reseproblem. Också Mammografienheten på ett sjukhus är en separat avdelning och bör inte påverka antal mängder patienter på ett sjukhus. Vi kunde inte fastställa kostnader för vare sig mobila eller fasta mammografienheter under denna studie. Därmed kan slutsatsen inte avgöras om det är förmånligt med mobila mammografienheter eller inte. Men däremot är en sak säker, det kommer att behövas att anställa fler personal.

Reinforcement Learning has applications in various domains, but the typical assumption is of a stationary process. Hence, when this hypothesis does not hold, performance may be sub-optimal. Tackling non-stationarity is not a trivial task because it requires adaptation to changing environments and predictability in various conditions, as dynamics and rewards might change over time. Meta Reinforcement Learning has been used to handle the non-stationary evolution of the environment while knowing the potential source of noise in the system. However, our research presents a novel method to manage such complexity by learning a suitable latent representation that captures relevant patterns for decision-making, improving the policy optimization procedure. We present a two-step framework that combines the unsupervised training of Deep Variational Auto-encoders to extract latent variables and a state-of-the-art model-free and off-policy Batch Reinforcement Learning algorithm called Fitted Q-Iteration, without relying on any assumptions about the environment dynamics. This framework is named Latent-Variable Fitted Q-Iteration (LV-FQI). Furthermore, to validate the generalization and robustness capabilities for exploiting the structure of the temporal sequence of time-series data and extracting near-optimal policies, we evaluated the performance with empirical experiments on synthetic data generated from classical financial models. We also tested it on Foreign Exchange trading scenarios with various degrees of non-stationarity and low signal-to-noise ratios. The results showed performance improvements compared to existing algorithms, indicating great promise for addressing the long-standing challenges of Continual Reinforcement Learning. / Reinforcement Learning har tillämpningar inom olika områden, men den typiska antagningen är att det rör sig om en stationär process. När detta antagande inte stämmer kan prestationen bli suboptimal. Att hantera icke-stationaritet är ingen enkel uppgift eftersom det kräver anpassning till föränderliga miljöer och förutsägbarhet under olika förhållanden, då dynamiken och belöningarna kan förändras över tiden. Meta Reinforcement Learning har använts för att hantera den icke-stationära utvecklingen av miljön genom att känna till potentiella källor till brus i systemet. Vår forskning presenterar emellertid en ny metod för att hantera en sådan komplexitet genom att lära en lämplig latent representation som fångar relevanta mönster för beslutsfattande och förbättrar optimeringsprocessen för policyn. Vi presenterar en tvåstegsramverk som kombinerar osuperviserad träning av Deep Variational Auto-encoders för att extrahera latenta variabler och en state-of-the-art model-free och off-policy Batch Reinforcement Learning-algoritm, Fitted Q-Iteration, utan att förlita sig på några antaganden om miljöns dynamik. Detta ramverk kallas Latent-Variable Fitted Q-Iteration (LV-FQI). För att validera generaliserings- och robusthetsförmågan att utnyttja strukturen hos den tidsmässiga sekvensen av tidsseriedata och extrahera nära-optimala policys utvärderade vi prestandan med empiriska experiment på syntetiska data genererade från klassiska finansiella modeller. Vi testade också det på handelsscenario för Foreign Exchange med olika grader av icke-stationaritet och låg signal-till-brus-förhållande. Resultaten visade prestandaförbättringar jämfört med befintliga algoritmer och indikerar stor potential för att tackla de långvariga utmaningarna inom kontinuerlig Reinforcement Learning.

This thesis focuses on visual multiple-target tracking (MTT) from a UAV. Typical state-of-the-art multiple-target trackers rely on an object detector as the primary detection source. However, object detectors usually require a GPU to process images in real-time, which may not be feasible to carry on-board a UAV. Additionally, they often do not produce consistent detections for small objects typical of UAV imagery.In our method, we instead detect motion to identify objects of interest in the scene. We detect motion at corners in the image using optical flow. We also track points long-term to continue tracking stopped objects. Since our motion detection algorithm generates multiple detections at each time-step, we use a hybrid probabilistic data association filter combined with a single iteration of expectation maximization to improve tracking accuracy.We also present a motion detection algorithm that accounts for parallax in non-planar UAV imagery. We use the essential matrix to distinguish between true object motion and apparent object motion due to parallax. Instead of calculating the essential matrix directly, which can be time-consuming, we design a new algorithm that optimizes the rotation and translation between frames. This new algorithm requires only 4 ms instead of 47 ms per frame of the video sequence.We demonstrate the performance of these algorithms on video data. These algorithms are shown to improve tracking accuracy, reliability, and speed. All these contributions are capable of running in real-time without a GPU.

<p dir="ltr">Employee staffing and scheduling are critical aspects of resource management in labor-intensive, customer-centric service organizations. This thesis investigates the optimal decision-making process for these critical tasks in the presence of non-stationary uncertainty, such as case-mix resident need, recommended staffing hours, and potential staffing turnover, a challenge prevalent in various domains, including healthcare and nursing home management.</p><p dir="ltr">The research begins predicting resident needs accurately. For this purpose, we present a novel Bayesian modeling approach to predict nursing home need-based resident census and staffing time. The resultant time series data of need-based resident census and staffing time are nonstationary with potential correlations between resource utilization groups. We thus propose Bayesian latent variable models with time-varying latent states to capture the dynamic patterns of resident service needs. We demonstrate the superiority of the proposed Bayesian prediction models by comparing their forecasting performance with several popular benchmark models, using historical assessment and aggregate staffing data from representative nursing homes.</p><p dir="ltr">The thesis further incorporates a rolling-horizon scheduling approach that integrates a periodically evolving Bayesian forecasting method into a series of stochastic look-ahead decision actions over multiple periods. To deal with the workforce scheduling with nonstationary demand uncertainty, we introduce a stochastic lookahead optimization framework that executes two-stage stochastic programming periodically along a rolling horizon to address the evolving non-stationary uncertainty. We obtain two-stage stochastic programming models to design effective work schedules, specifically assigning nurses to various shifts while balancing the staff workload and accommodating fluctuating resident needs.</p><p dir="ltr">We finally introduce the SNHSSO framework (stochastic nursing home staffing and scheduling optimizer), encompassing data modeling and addressing multi-period, multi-uncertainty, and multi-objective staffing and scheduling challenges. When the SNHSSO Optimizer is executed with the provided inputs, it generates recommended staffing decisions for longer planning horizons, as well as schedules and contingency plans for shorter planning horizons. These adapted decisions and adjusted parameters are archived for future reference, facilitating subsequent iterations of the process. SNHSSO optimizes caregiver assignments by taking into account probabilistic forecasts of service requirements, resident acuity, and staff turnover, all within two-stage stochastic mixed integer linear programs. Our approach leverages a scenario-based rolling horizon methodology to effectively solve the SNHSSO model.</p><p dir="ltr">The empirical foundation of this work is built on case studies conducted using Minimum Data Set (MDS) data spanning five years from 2014 to 2018 in Indiana nursing homes.</p>

Performing causal reasoning in a scene is an inherent mechanism in human cognition; however, the majority of approaches in the causality literature aiming for this task still consider constrained scenarios, such as simple physical systems or stationary time-series data. In this work we aim for causal discovery in videos concerning realistic scenarios. We gather motivation for causal discovery by acknowledging this task to be core at human cognition. Moreover, we interpret the scene as a composition of time-series that interact along the sequence and aim for modeling the non-stationary behaviors in a scene. We propose State-dependent Causal Inference (SDCI) for causal discovery in conditional stationary time-series data. We formulate our problem of causal analysis by considering that the stationarity of the time-series is conditioned on a categorical variable, which we call state. Results show that the probabilistic implementation proposed achieves outstanding results in identifying causal relations on simulated data. When considering the state being independent from the dynamics, our method maintains decent accuracy levels of edge-type identification achieving 74.87% test accuracy when considering a total of 8 states. Furthermore, our method correctly handles regimes where the state variable undergoes complex transitions and is dependent on the dynamics of the scene, achieving 79.21% accuracy in identifying the causal interactions. We consider this work to be an important contribution towards causal discovery in videos. / Att utföra kausala resonemang i en scen är en medfödd mekanism i mänsklig kognition; dock betraktar fortfarande majoriteten av tillvägagångssätt i kausalitetslitteraturen, som syftar till denna uppgift, begränsade scenarier såsom enkla fysiska system eller stationära tidsseriedata. I detta arbete strävar vi efter kausal upptäckt i videor om realistiska scenarier. Vi samlar motivation för kausal upptäckt genom att erkänna att denna uppgift är kärnan i mänsklig kognition. Dessutom tolkar vi scenen som en komposition av tidsserier som interagerar längs sekvensen och syftar till att modellera det icke-stationära beteendet i en scen. Vi föreslår Tillståndsberoende kausal inferens (SDCI) för kausal upptäckt i villkorlig stationär tidsseriedata. Vi formulerar vårt problem med kausalanalys genom att anse att tidsseriens stationäritet är villkorad av en kategorisk variabel, som vi kallar tillstånd. Resultaten visar att det föreslagna probabilistiska genomförandet uppnår enastående resultat vid identifiering av orsakssambandet på simulerade data. När man överväger att tillståndet är oberoende av dynamiken, upprätthåller vår metod anständiga noggrannhetsnivåer av kanttypsidentifiering som uppnår 74, 87% testnoggrannhet när man överväger totalt 8 tillstånd. Dessutom hanterar vår metod korrekt regimer där tillståndsvariabeln genomgår komplexa övergångar och är beroende av dynamiken på scenen och uppnår 79, 21% noggrannhet för att identifiera kausala interaktioner. Vi anser att detta arbete är ett viktigt bidrag till kausal upptäckt i videor.

No description available.

Power outages, electric-grid deficiencies and renewable energies are all examples where stationary energy storages are useful. In this master thesis, two types of stationary electrochemical energy storages are examined; vent-regulated lead-acid batteries (VRLA) and lithium iron phosphate batteries (LFP), to find out the more beneficial one in stationary uses. The techniques are examined for a large range of electric-grid services in a techno-economical investigation. The cost per delivered kWh of the energy storage is the basis of comparison which is calculated using battery degradation data with respect to C-rate, SoC, DoD, temperature, storage time and cycle frequency to estimate calendar and cyclic aging. Modelling presents neither alternative as superior although LFP is the more versatile alternative. VRLA-batteries can be a more cost-beneficial alternative for applications demanding less than 1 cycle/day, at temperatures lower than 30C, short project lifetimes and when utilizing storages beyond 80% EoL. The investment cost is lower for VRLA at equal C-rates. Cost items neglected will decrease the chances of VRLA being the cheapest technique. From a sustainability point of view, LFP is under almost all circumstances the less energy and CO2-intense technology, however recyclability is in clear favor for VRLA. / Strömavbrott, underdimensionerade elnät och förnybar energi är tre exempel där ett stationärt energilager kan tillämpas. I den här masteruppsatsen undersöks två typer av stationära elektrokemiska energilager; ventilreglerade bly-syra-batterier och litium-järnfosfat-batterier (LFP), för att finna det mer fördelaktiga alternativet i stationära applikationer. De två teknikerna analyseras i ett stort antal elnätsapplikationer i en tekno-ekonomisk studie. Kostnaden per levererad kWh av energilagret används som jämförelsebas vilken beräknas utifrån batteridegraderingsdata med avseende på C-rate, SoC, DoD, temperatur, lagringstid och cykelfrekvens för att estimera kalender- och cyklisk åldring. Modellering visar att inget av batterialternativen är överlägset i alla situationer men LFP är det mångsidigare alternativet. Bly-syra-batterier kan vara mer kostnadseffektiva för applikationer som kräver mindre än 1 (full-ekvivalent) cykel/dag vid temperaturer lägre än 30C, korta projektlivstider samt när batterilagren används bortom 80% EoL. Investeringskostnaden är lägre för bly-syra-batterier när likadan C-rate appliceras. Negligerade kostnadsposter kommer minska chanserna att bly-syra-batterier är det billigaste alternativet. Från ett hållbarhetsperspektiv är LFP nästan uteslutande den mindre energikrävande och mindre CO2-intensiva tekniken. Bly-syra-batterier har dock en klar fördel när det kommer till återvinningsbarhet.

The electrification of heavy transport is considered a solution to climate change. Today, electrification is challenged by high investment costs for vehicles and charging, which hampers the conversion. The purpose is to investigate what investment risks occur when creating a stationary charging hub by focusing on three different charging possibilities (private, semi-public, and public) to examine the differences between them. The study is a case study with the study object Stockholm South, which is considered to be a future logistics area. The study is based on literature, documents, and interviews with actors linked to the study object. Moreover, to identify potential investment risks the study conducts a life cycle cost analysis of a stationary charging hub. The conducted interviews identified that several actors are positive to enable a charging hub, but are challenged by high investment costs, uncertainties regarding electricity network capacity, and future prospects linked to the technological development of electric trucks and charging technology. Moreover, the interviews also identified opportunities such as financial support, potential market advantages of being one of the firsts on the market, and contribution to a sustainable transport sector. The results of the cost analysis show that a stationary charging hub for heavy goods transport entails high investments and operating costs, which are affected by future aspects such as electricity prices, electricity networks, technology development, and utilisation rate. In addition, the choice of closed or open charging hub has an impact on the total cost. The analysis of the results indicates that collaboration between the involved actors can facilitate several of the identified challenges and risks. The study contributes to expanding the empirical data on stationary charging infrastructure for heavy freight transport. In addition, the study contributes with a cost analysis for a charging station that can provide inspiration and a basis for future investments. The theoretical contribution mainly concerns how new technologies in ecosystem theory are handled in relation to collaboration and business models. / Elektrifieringen av tunga transporter anses vara en lösning på klimatförändringarna. Elektrifieringen utmanas idag av höga investeringskostnader för fordon och laddning, vilket hämmar omställningen. Syftet med studien är att undersöka vilka investeringsrisker som uppstår vid skapandet av en stationär laddhubb genom att fokusera på tre olika laddningsmöjligheter (privat, semi-publik och publik) för att undersöka skillnaderna mellan dem. Studien är en fallstudie där studieobjektet är Stockholm Syd som anses vara ett framtida logistikområde. Studien baseras på litteratur, dokument och intervjuer med aktörer kopplat till studieobjektet. För att kunna identifiera investeringsriskerna har en livscykelkostnadsanalys gjorts för en stationär laddhubb. De genomförda intervjuerna identifierade att flera aktörer är positiva till att skapa en ladd hubbmen att utmaningarna framförallt ligger i höga investeringskostnader, osäkerheter angående elnätskapacitet och framtidsutsikter kopplat till den teknologiska utvecklingen av ellastbilar och laddningsteknik. Däremot identifieras även möjligheter så som finansiella stöd, potentiella marknadsfördelar med att vara en av de första på marknaden och att bidra till en hållbar utveckling av transportsektorn. Resultatet av kostnadsanalysen visar att en stationär laddhubb för tunga godstransporter medför höga investerings- och operationskostnader som påverkas av framtida aspekter så som elpris, elnät, teknikutveckling samt utnyttjandegrad. Dessutom har val av stängd eller öppen laddstation inverkan på kostnaderna. Analys av resultatet tyder på att samarbete mellan involverade aktörer kan underlätta flera av de identifierade utmaningarna och riskerna. Studien bidrar till en utökad empiri kring stationär laddinfrastruktur för tunga godstransporter. Dessutom bidrar studien med en kostnadsanalys för en laddstation som kan ge inspiration och underlag för framtida investeringar. Det teoretiska bidraget berör främst hur nya tekniker ses på inom ekosystemsteorin i förhållande till samarbete och affärsmodeller.

Digital Signal Processing has been dominated by the Fourier transform since the Fast Fourier Transform (FFT) was developed in 1965 by Cooley and Tukey. In the 1980's a new transform was developed called the wavelet transform, even though the first wavelet goes back to 1910. With the Fourier transform, all information about localized changes in signal features are spread out across the entire signal space, making local features global in scope. Wavelets are able to retain localized information about the signal by applying a function of a limited duration, also called a wavelet, to the signal. As with the Fourier transform, the discrete wavelet transform has an inverse transform, which allows us to make changes in a signal in the wavelet domain and then transform it back in the time domain. In this thesis, we have investigated the filtering properties of this technique and analyzed its performance under various settings. Another popular application of wavelet transform is data compression, such as described in the JPEG 2000 standard and compressed digital storage of fingerprints developed by the FBI. Previous work on filtering has focused on the discrete wavelet transform. Here, we extended that method to the stationary wavelet transform and found that it gives a performance boost of as much as 9 dB over that of the discrete wavelet transform. We also found that the SNR of noise filtering decreases as a frequency of the base signal increases up to the Nyquist limit for both the discrete and stationary wavelet transforms. Besides filtering the signal, the discrete wavelet transform can also be used to estimate the standard deviation of the white noise present in the signal. We extended the developed estimator for the discrete wavelet transform to the stationary wavelet transform. As with filtering, it is found that the quality of the estimate decreases as the frequency of the base signal increases. Many interesting signals are self-similar, which means that one of their properties is invariant on many different scales. One popular example is strict self-similarity, where an exact copy of a signal is replicated on many scales, but the most common property is statistical self-similarity, where a random segment of a signal is replicated on many different scales. In this work, we investigated wavelet-based methods to detect statistical self-similarities in a signal and their performance on various types of self-similar signals. Specifically, we found that the quality of the estimate depends on the type of the units of the signal being investigated for low Hurst exponent and on the type of edge padding being used for high Hurst exponent. / Master of Science

We set out to analyze 5D stationary and bi-axisymmetric solutions to the vacuum Einstein equations. These are in the cohomogeneity 2 setting where the orbit space is a right half plane. They can have a wide range of behaviour at the boundary of the orbit space. The goal is to understand in detail the soliton example in Khuri, Weinstein and Yamada's paper ``5-dimensional space-periodic solutions of the static vacuum Einstein equations". This example is periodic and has alternating axis rods as its boundary data. We start by deriving the harmonic equations which determines the behaviour of the metric in the interior of the orbit space. Then we analyze what conditions the boundary data imposes on the metric. These are called the smoothness conditions which we derive for solely the alternating axis rod case. We show that with an ellipticity assumption they predict that the twist potentials are constant and that the metric is of the form which appears in Khuri, Weinstein and Yamada's paper. We then analyze the Schwarzschild metric in its standard form which is cohomogeneity 1 and its Weyl form which is cohomogeneity 2. This Weyl form can be made periodic and this serves as an inspiration for the examples in Khuri, Weinstein and Yamada's paper. Finally we analyze the soliton example in detail and show that it satisfies the smoothness conditions. We then provide a new example which has a single axis rod on the boundary with non-constant twist potentials but that is missing a point on the boundary. / Thesis / Master of Science (MSc) / We study the geometry of 5D blackholes. These blackholes are idealized by certain spatial symmetries and time invariance. They are solutions to the vacuum Einstein equations. The unique characteristic of these blackholes is the range of behaviour they may exhibit at the boundary of the domain of outer communication. There could be a standard event horizon called a horizon rod or an axis rod where a certain part of the spatial symmetry becomes trivial. In this thesis we start by deriving the harmonic map equations which are satisfied in the interior of the domain of communication. Then we show how this boundary data affects the metric through the smoothness conditions. We then analyze the soliton example in a paper by Khuri, Weinstein and Yamada and show that it respects the smoothness conditions. We then provide a new example which is interesting in the fact it has non-constant twist potentials.