As alterações que a ação antrópica vem causando nos ambientes costeiros tem provocado impactos sobre as espécies a eles associadas. Larus dominicanus é uma espécie de ave marinha amplamente distribuída ao longo do Hemisfério Sul. Por possuir um hábito alimentar generalista, os resíduos da ação antrópica tem beneficiado a espécie, que, assim como outras gaivotas, vem apresentando um crescimento demográfico acelerado. O presente estudo, através do uso de marcadores de microssatélites, mostra que, apesar disso, a espécie possui uma baixa variabilidade genética, com fraca estruturação populacional, que provavelmente são o reflexo da origem recente da espécie e de uma diferenciação recente entre as populações. Múltiplas forças atuam para determinar a estruturação populacional, sendo elas o isolamento por distância, as barreiras físicas e a filopatria. Os sinais de gargalo populacional encontrados em algumas das colônias levantam a possibilidade de efeitos fundadores por colonização recente nas colônias mais ao norte da costa brasileira e redução populacional nas colônias da Argentina e da Antártica como conseqüência da última glaciação. Estes dados chamam a atenção para a necessidade de considerar as informações genéticas para a implantação de planos de manejo. Uma vez que a diferenciação entre as populações é recente, a variabilidade dentro de cada uma delas deve ser mantida. O controle populacional da espécie através de métodos diretos deve ser também acompanhado por planos de manejo ambiental, visando reduzir ou eliminar as condições que propiciam o crescimento desequilibrado dos gaivotões. / The changes in the coast that has been caused by human action has led to impacts on species associated with this environment. Larus dominicanus is a seabird species widely distributed throughout the Southern Hemisphere. The generalist feeding habits allow this species take advantage from human action, leading to population growth, which is also observed in other species of gulls. This study, through the use of microsatellite markers, shows that, despite of the population growth, Larus dominicanus has a low genetic variability, with low population structure, which probably reflects the recent origin of species and a recent differentiation among populations. Multiple forces act to determine the population structure, among them the isolation by distance, physical barriers and philopatry. Some colonies presents a bottleneck sign, raising the hypothesis of recent founder effects in the colonies to the north of Brazil and population reduction of colonies of Argentine and Antarctic as consequence of the last glaciation. These data show the need to consider genetic information for the implementation of management plans. The variability within populations must be maintained, since the differentiation between them is recent. Furthermore, the species population control by direct methods must also be accompanied by environmental management plans, to reduce or eliminate the conditions that favor the unbalanced growth of the gulls

Process mining (PM) has already found its place in analyzing and visualizing business processes across different sectors such as banking, healthcare, insurance, and logistics. However, its application is very limited in production systems. Production systems are very complex and dynamic, hindering the accurate capture of different process chain events. Generating the event log data out of the recorded events requires a proper understanding of the processes. The results of the bottleneck analysis can help the production experts to optimize the process. This study analyzes three to four PM tools to select an appropriate tool that can perform bottleneck analysis. The initial research will focus on various features of the PM tools such as modeling support, user friendliness, flexibility, stability, animation support, and BPMN modeling support. The feature analysis uses the event-log data generated from simulating the process model created in a Simulation-based Optimization tool FACTS Analyzer. Some PM tools can generate the BPMN model from the process model; animation support and modeling support features are vital for the bottleneck analysis. The PM tool's ability to create the process model automatically from the event log data can help save time involved in data collection through a physical visit to the site, conducting interviews, observations, etc. In contrast, the bottleneck analysis used simulated event log data and an event log generated from a real-world production system. The study focuses on finding whether PM can support the bottleneck analysis of the real-world production systems and find the appropriate PM tool available in the market for the same application. The use of process mining tools in the bottleneck analysis requires human intervention in identifying the bottlenecks in the process as the PM tools are incapable of recognizing them on their own. But, appropriate use of the PM tools in the bottleneck analysis can help save a considerable amount of time in manually building the simulation model. Performing iterative optimization could lead to socio, economic, and environmental sustainability as it helps avoid wastages in time, material, and energy.

The topic of diploma thesis was implemented in selected mechanical engineering company producing precision parts in small series. The main aim of the thesis is to create the suggestion of optimal supply process for new orders. The first part of thesis describes the company, area of its business and the processes of the order. In second part, the production analysis is carried out: production order path, production analysis for two reference parts, comparison of all time delay over two last years, find bottlenecks in the supply process. Base on analysis, theoretical approaches was described. In the last chapter there are suggestions for improvement of supply processes and thus shortening of production time of new orders.

The competition between manufacturing companies is constantlyincreasing. There is therefore a need for continuously improvingproduction control and production planning. The requirements arepartly founded on higher customer expectations on quality anddelivery precision. These requirements make manufacturing operationsmore demanding. To become more competitive, companies often turn tovirtual engineering tools with the intention of finding improvementsproposals that are not always completely obvious. The aim of this thesis work was to identify bottlenecks that limitsystem line production with the help of discrete event simulationand to identify improvement proposals that were able to reduce theproduction lead time. The work has been carried out at ABB inLudvika at the department for manufacturing the outer casing forpower transformers. The reliability of the study was ensured byusing a mixture of qualitative and quantitative data collectionmethods that complemented each other appropriately. Process mappingand collected data were the bases a simulation model, which is thebasis for the achieved work result. Analysis of the simulation results gave strong indications that aspecific operation in the production line needed to be improved andthe need to better balance operator capacity within the line. Theprocess mapping and data analysis highligthted the need for thecompany to work more with process standardisation. The finalimprovement proposals submitted to the company are intended toincrease the delivery precision and production capacity / Konkurrensen mellan tillverkande företag ökar ständigt, behovet avkontinuerlig effektivisering av produktionsstyrning och planering ärsåledes ett faktum. Tillverkning försvåras samtidigt då kunderförväntar sig allt mer kvalitativa och unika produkter med högleveransprecision. För att bli mer konkurrenskraftig vänder sig oftaföretag till virtuella hjälpmedel med avseendet att finnaförbättringsförslag som inte alltid är helt självklara. Syftet med arbetet var att identifiera eventuella flaskhalsar sombegränsar produktionen med hjälp av händelsestyrd simulering och attidentifiera förbättringsförslag som påverkar produktionskapaciteten.Arbetet har utförts på ABB i Ludvika vid avdelningen förtillverkning av ytterhöljet till transformatorer. Studienstillförlitlighet säkerställdes genom användandet av en blandningmellan kvalitativa och kvantitativa datainsamlingsmetoder somkompletterade varandra. Processkartläggning och datainsamling vargrunden för att ta fram en valid simuleringsmodell som var grund fördet uppnådda arbetsresultatet. Analys av simuleringsresultatet gav starka indikationer på att enspecifik operation i produktionen behövde justeras samt attföretaget behöver jobba med balansering av operatörs resurser föratt bättre möta kapacitetsbehovet. Utöver detta finns det ettfortsatt behov att arbeta med process standardisering. Deslutgiltiga förbättringsförslagen som lämnats till företaget ärvägledande för att öka leveransprecisionen ochproduktionskapaciteten.

This project was implemented during spring 2021 as a bachelor thesis for the Bachelor Engineering Program in Mechanical Engineering at Karlstads University. The project wascarried out in collaboration with Stora Enso Skoghall Mill, which produces different types of cardboard material. In 2020, the overall production capacity was 875 000 tons of cardboard material. In November 2020 Skoghall Mill was authorized to precede an investigation regarding a possible production increase, which would result in a self-sufficient pulp production and an increase of 120 000 tons cardboard material per year.  The cardboard is winded into rolls before they are being transported to the customers. Before the transportation, the reels are given a protecting wrapping. This procedure occurs in a production segment called the Wrapping Line, which is a part of the Finishing Department. The Wrapping Line is one of the production areas that would be affected by a potential increase of production. Therefore, the responsible of the unit wanted to examine if the Wrapping Linehad enough capacity for a potential increase of this size.  The purpose of this paper was to carry out a capacity analysis for the line, identify bottlenecks in the flow and to present arrangements to make the production line more effective. Furthermore, an additional purpose was to identify how long the maintenance stops could beif the potential production increase is approved.  The capacity analysis was carried out using a simulation software called JaamSim. A model of the existing system was built in JaamSim, which afterwards was analyzed. The capacity analysis showed that Emballeringen has the needed capacity if the production increase isapproved. The capacity analysis also displayed the possible length of maintenance stops. Arrangements in flow was presented and prioritized so that the capacity in the Wrapping Linewould increase further.

<p>In this work, we study a class of Markovian information theoretic optimization problems motivated by the recent interests in incorporating mutual information as performance metrics which gives evident success in representation learning, feature extraction and clustering problems. In particular, we focus on the information bottleneck (IB) and privacy funnel (PF) methods and their recent multi-view, multi-source generalizations that gain attention because the performance significantly improved with multi-view, multi-source data. Nonetheless, the generalized problems challenge existing IB and PF solves in terms of the complexity and their abilities to tackle large-scale data. </p> <p>To address this, we study both the IB and PF under a unified framework and propose solving it through splitting methods, including renowned algorithms such as alternating directional method of multiplier (ADMM), Peaceman-Rachford splitting (PRS) and Douglas-Rachford splitting (DRS) as special cases. Our convergence analysis and the locally linear rate of convergence results give rise to new splitting method based IB and PF solvers that can be easily generalized to multi-view IB, multi-source PF. We implement the proposed methods with gradient descent and empirically evaluate the new solvers in both synthetic and real-world datasets. Our numerical results demonstrate improved performance over the state-of-the-art approach with significant reduction in complexity. Furthermore, we consider the practical scenario where there is distribution mismatch between training and testing data generating processes under a known bounded divergence constraint. In analyzing the generalization error, we develop new techniques inspired by the input-output mutual information approach and tighten the existing generalization error bounds.</p>

Presented PhD thesis aims at development of new methodologies and approaches to reduce lead times in production systems. Therefore the thesis examines in detail the design of new, more comprehensive approaches to shortening lead times. The dissertation describes the methodology of using these approaches in production systems, it is suggested their inclusion in the control structure of production and adequate managerial support for the successful application of these new approaches in the industry is developed. The first part summarizes current knowledge in the field of reducing lead times. The various methods that can be used for reducing lead times are defined and described, from which is selected discrete simulation of manufacturing systems as a suitable method for the purposes of issue development. The next part of this work is therefore devoted to simulation approach to reducing lead times and the factors that affect them, such as processing time, size of the batch, queue lengths, etc. In this part is also discussed the use of information technologies (programming languages, libraries, and software tools) that enable efficient creation of simulation models. The second part is focused on industrial applications implemented to reduce production lead times. Specific simulation analysis deal with, for example, reduction of the lead time for manufacture of surgical instruments, screws, slats, but also transport systems, optimization of production orders input to the production system, the issue of supply, etc. The second part also describes a practical approach to data transfer to / from the simulation model, simulation software connection to the ERP database of enterprise and energy intensity of production. The conclusion summarizes the findings that were detected on the issue of reducing lead times, and within the chosen methodology, simulation of manufacturing processes indicates possible further progress in the research and practical applications.

I det här arbetet eftersökte Scania IM en metod för att beräkna maskiners påverkan på output för en produktionslina vid driftstörningar. Metoden skulle kunna appliceras på en produktionslina med ett serie-parallellt flöde och arbetet utgick från en utvald lina där man bearbetar cylinderhuvuden. Vid framtagandet av beräkningsmetoden skapades det som i arbetet kallas för påverkansfaktor som tar fram vilken påverkan ett bortfall av en maskin i flödet har på flaskhalsutrustningen, som i sin tur står för den yttersta begränsningen för systemet. På seriella stationer erhålls en påverkansfaktor på 100 % men på stationer med parallell utrustning så finns det en kvarvarande kapacitet vid driftstopp av enstaka maskiner vilket är beroende av hur många maskiner samt vilken utnyttjandegrad de hade i ursprungsläget. Vid framtagandet av beräkningsmetoden togs inspiration från tidigare arbeten inom detta område men anpassades eller förenklades för att möta behoven från underhållsavdelningen. En utveckling av metoden gjordes även för att kunna analysera historiska data där information omproducerad variant saknas och då används ett viktat genomsnitt för påverkansfaktorn. Påverkansfaktorn bortser från buffertars påverkan på systemet men hur de ska inkorporeras undersöktes och därav togs ett förslag fram där kritisk driftstoppstid (critical downtime) används tillsammans med påverkansfaktorn för att ta hänsyn till såväl buffertar som parallell utrustning. Ett verktyg skapades för Scania IM för att underlätta användandet av framtagna beräkningsmetoder. Det som kvarstår efter avslutat arbete är att över tid testa de framtagna metoderna för att undersöka hur väl de överensstämmer med verkligt produktionsutfall. Detta kunde ej genomföras under gällande situation då produktionen ej var i drift enligt normalläge under tidsperioden för arbetet. / For this thesis, Scania IM sought a method for calculating the impact of machine downtime on the output of a production line. The method was created to be applied on a production line with a seriesparallel flow and was constructed based on a selected line where cylinder heads are machined. To achieve this a variable called impact factor was created that produces the effect an individual machines downtime has on the bottleneck equipment for the production line, which in turn accounts for the utmost limitation of the system. At serial stations, an impact factor of 100% is obtained, but at stations with parallel equipment, there is a remaining capacity if at least one machine is functioning. The amount on remaining capacity is dependent on how many of the parallel machines that are malfunctioning and the degree of utilization they had in the original state. In developing the calculation method, inspiration was taken from previous work regarding similar subjects, but it was either adapted or simplified to meet the needs of the maintenance department for the production line. A further development of the method was also made to be able to analyze historical data where information on the produced variant of cylinder heads is missing which led to the use of a weighted average for the impact factor. The impact factor ignores the effects of buffers on the system output, but how they are to be incorporated was investigated and a proposal was therefore made where critical downtime is used together with the impact factor to take into account both buffers and parallel equipment. A tool was created for Scania IM where they can perform these calculations. What remains after completion of the thesis is to test the developed methods over time to examine how well they correspond to actual production data. This could not be done during the time of this thesis due to the current situation when production was not operating according to normal conditions.

Manufacturing companies throughout the world are interested in reducing the time between a customer placing an order and them receiving the payment for that order. This premise is something that is a central characteristic for the Lean philosophy, and is one of the reasons to apply it. Today manufacturers around the world are embracing Lean techniques in order to reduce waste and increase productivity, and also increase the inventory turns, which reflects in an improvement of cash flow for the company. Nowadays, with all the financial turmoil, every company is looking forward to reduce the inventories, to work with Just in Time supply chains, to develop production systems that reduce the scrap and produce only what is needed, saving space, and freeing up time to work on new design and be at the edge of innovation in order to gain market share and keep improving. This master thesis is focused on implementing the Lean principles in a pillow production line, in order to achieve it, a series of techniques to assess the facility where implemented, which allowed to understand how the facility was working, where is the bottleneck, and to understand the function of it as a system, avoiding to focus on a single point but viewing it as a whole, where each part contributes in a specific and unique way, but where all of them are necessary. Applying Lean principles is a daunting task that takes a long time, a never ending trial and error process, because of this the goal of this study is to develop the bases for a Lean transformation, a schedule for the implementation will be developed and proposed to the company, after analyzing the facility. The study reveals that it is possible to reduce the lead-time of the facility in 60%, and avoid the backorders situation that is present in the company, improving also the service level.

Although deep neural networks (DNNs) have made remarkable achievementsin various elds, there is still not a matching practical theory that is able toexplain DNNs' performances. Tishby (2015) proposed a new insight to analyzeDNN via the Information bottleneck (IB) method. By visualizing how muchrelevant information each layer contains in input and output, he claimed thatthe DNNs training is composed of tting phase and compression phase. Thetting phase is when DNNs learn information both in input and output, andthe prediction accuracy goes high during this process. Afterwards, it is thecompression phase when information in output is preserved while unrelatedinformation in input is thrown away in hidden layers. This is a tradeo betweenthe network complexity (complicated DNNs lose less information in input) andprediction accuracy, which is the same goal with the IB method.In this thesis, we verify this IB interpretation rst by reimplementing Tishby'swork, where the hidden layer distribution is approximated by the histogram(binning). Additionally, we introduce various mutual information estimationmethods like kernel density estimators. Based upon simulation results, we concludethat there exists an optimal bound on the mutual information betweenhidden layers with input and output. But the compression mainly occurs whenthe activation function is \double saturated", like hyperbolic tangent function.Furthermore, we extend the work to the simulated wireless model where thedata set is generated by a wireless system simulator. The results reveal that theIB interpretation is true, but the binning is not a correct tool to approximatehidden layer distributions. The ndings of this thesis reect the informationvariations in each layer during the training, which might contribute to selectingtransmission parameter congurations in each frame in wireless communicationsystems. / Ä ven om djupa neuronnät (DNN) har gjort anmärkningsvärda framsteg på olikaområden, finns det fortfarande ingen matchande praktisk teori som kan förklara DNNs prestanda. Tishby (2015) föreslog en ny insikt att analysera DNN via informationsflaskhack (IB) -metoden. Genom att visualisera hur mycket relevant information varje lager innehåller i ingång och utgång, hävdade han att DNNs träning består av monteringsfas och kompressionsfas. Monteringsfasenär när DNN lär sig information både i ingång och utgång, och prediktionsnoggrannheten ökar under denna process. Efteråt är det kompressionsfasen när information i utgången bevaras medan orelaterad information i ingången kastas bort. Det här är en kompromiss mellan nätkomplexiteten (komplicerade DNN förlorar mindre information i inmatning) och predictionsnoggrannhet, vilket är exakt samma mål med informationsflaskhals (IB) -metoden.I detta examensarbete kontrollerar vi denna IB-framställning först genom att implementera om Tishby’s arbete, där den dolda lagerfördelningen approximeras av histogrammet (binning). Dessutom introducerar vi olika metoder förömsesidig information uppskattning som kernel density estimators. Baserat på simuleringsresultatet drar vi slutsatsen att det finns en optimal bindning för denömsesidiga informationen mellan dolda lager med ingång och utgång. Men komprimeringen sker huvudsakligen när aktiveringsfunktionen är “dubbelmättad”, som hyperbolisk tangentfunktion.Dessutom utvidgar vi arbetet till den simulerad trådlösa modellen där data set genereras av en trådlös systemsimulator. Resultaten visar att IB-framställning är sann, men binningen är inte ett korrekt verktyg för att approximera dolda lagerfördelningar. Resultatet av denna examensarbete reflekterar informationsvariationerna i varje lager, vilket kan bidra till att välja överföringspa-rameterns konfigurationer i varje ram i trådlösa kommunikationssystem