Due to its ubiquitous availability, Short Message Service (SMS), first introduced in the 1980s, became not only the most popular way of communication, but also stimulated the development of SMS-based value added services. This application-to-person traffic is delivered to end users through SMS aggregators who provide the link between service providers and mobile carriers. In order to perform optimal traffic routing, the aggregators need to estimate the quality of each potential international route to the specified destination. The evaluation criteria include end-to-end delivery time, as well as correct verification of delivered data. This thesis suggests a method of quality of service (QoS) assessment for international SMS service which combines two types of tests, end-to-end delay measurements and various verification tests. A prototype of the testing system for international SMS service was developed to generate SMS traffic, collect and analyze results, and evaluate the experienced QoS of the SMS route used in accordance with the proposed approach. As a part of end-to- end delay measurement tests, SMS traffic was sent to Singtel network in Singapore along two routes. The verification tests were executed via different routes to two mobile networks: Singtel and Tele2 (Sweden). The results of the performed measurements determined the route with the highest QoS, i.e. the one with bigger bottleneck bandwidth and lower data loss rate. The prototype of the SMS testing system can be used by SMS aggregators to verify delivery of a SMS message, check the integrity of the message, figure out interconnection type of the route supplier with the destination carrier and to identify the presence of load balancers in the path. The prototype also makes it possible to compare end-to-end delay times of several routes and compute bottleneck values for each of the tested routes.

For different input contexts (sequence of previous words), to predict the next word, a neural word-level language model outputs a probability distribution over all the words in the vocabulary using a softmax function. When the log of probability outputs for all such contexts are stacked together, the resulting matrix is a log probability matrix which can be denoted as Q_theta, where theta denotes the model parameters. When language modeling is formulated as a matrix factorization problem, the matrix to be factorized Q_theta is expected to be high-rank as natural language is highly context-dependent. But existing softmax based word-level language models have a limitation of not being able to produce such matrices; this is known as the softmax bottleneck. There are several works that attempted to overcome the limitations introduced by softmax bottleneck, such as the models that can produce high-rank Q_theta. During the process of reproducing the results of these works, we observed that the rank of Q_theta does not always positively correlate with better performance (i.e., lower test perplexity). This puzzling observation triggered us to conduct a systematic investigation to check the influence of rank of Q_theta on better performance of a language model. We first introduce a new family of activation functions called the Generalized SigSoftmax (GSS). By controlling the parameters of GSS, we were able to construct language models that can produce Q_theta with diverse ranks (i.e., low, medium, and high ranks). For models that use GSS with different parameters, we observe that rank does not have a strong positive correlation with perplexity on the test data, reinforcing the support of our initial observation. By inspecting the top-5 predictions made by different models for a selected set of input contexts, we observe that a high-rank Q_theta does not guarantee a strong qualitative performance. Then, we conduct experiments to check if there are any other additional benefits in having models that can produce high-rank Q_theta. We expose that Q_theta rather suffers from the phenomenon of fast singular value decay. Additionally, we also propose an alternative metric to denote the rank of any matrix known as epsilon-effective rank, which can be useful to approximately quantify the singular value distribution when different values for epsilon are used. We conclude by showing that it is the regularization which has played a positive role in the performance of these high-rank models in comparison to the chosen baselines, and there is no single model yet which truly gains improved expressiveness just because of breaking the softmax bottleneck.

Dagens industrier utvecklas snabbare och ständigt uppstår nya teknikutvecklingar som skapar nya krav och kräver förändringar inom tillverkningsindustrin. Eliminering av flaskhalsar är viktigt inom läkemedelsindustrin för att möta den ökade efterfrågan. Med tanke på att läkemedelsindustrin växer har polymerase chain reaction-tekniken (PCR) blivit en av de mest värdefulla teknikerna inom biovetenskap, diagnostik och rättsmedicin. Utvecklingen av PCR-teknik drivs av flera marknadsaktörer och skapar hård konkurrens. På grund av den höga konkurrensen är det avgörande att främja teknisk utveckling och skapa en optimal produktion inom läkemedelsindustrin för att möta marknadens efterfrågan. Det kan vara kostsamt att producera pålitlig utrustning inom läkemedelsindustrin. För att skapa affärsmål och analysera förbättringsmöjligheter inom produktionsprocessen kan optimeringssystem skapas med hjälp av simulering. Genom att skapa en virtuell modell kan data samlas för att identifiera problem inom tillverkningsindustrin. Genom flödessimulering kan en flaskhalsanalys skapas med fokus på väntetider, cykeltider och genomströmning. Denna studie svarar på forskningsfrågorna: (1) Vilka problem identifierar flödessimuleringar vid analys av prestanda och flaskhalsar för förpackningslinjer inom läkemedelsindustrin? (2) Vilka effekter får lösningsförslaget på prestandan för flaskhalsflödet i produktionsprocessen? Genom att simulera olika scenarier och identifiera problem relaterade till resursanvändning och prioriteringar vid sammanfogningspunkter kan lösningar på flaskhalsar identifieras. Genom att jämna ut prioriteringarna och använda buffertlager kan företag skapa ett jämnare flöde och förbättra motståndskraften i sina tillverkningssystem. Studien visar att design- och layoutval har en betydande effekt på den totala genomströmningen, vilket betonar vikten av noggrann layoutplanering och optimering inom läkemedelsindustrin. Genom att förbättra nyckeltal som medeltid mellan fel och medeltid mellan reparation kan företag öka genomströmning, produktivitet och effektivitet samtidigt som de uppnår ekonomisk, ekologisk och social hållbarhet. Rekommendationer för framtida studier inkluderar att undersöka operatörsplacering och maskinhastighetens inverkan på väntetiden. / Today's industries are developing at a faster pace and new technological developments are constantly being developed that create new demands and require changes within the manufacturing industry. Eliminating bottlenecks is important in the pharmaceutical industry in order to meet the increased demand. Given that the pharmaceutical industry is a growing industry, polymerase chain reaction (PCR) technology has become one of the most valuable technologies used in life sciences, diagnostics and forensics. The development of PCR technology is carried out by several market players and causes strong competition. Due to the high level of competition, it is essential to promote technological development and create optimal production in the pharmaceutical industry in order to meet market demand. It can be expensive to produce reliable equipment in the pharmaceutical industry. To create business goals and analyze improvement opportunities within the production process, optimization systems can be created using simulation. By creating a virtual model, data can be collected to identify problems within the manufacturing industry. With the help of flow simulation, a bottleneck analysis can be created with a focus on waiting times, cycle times and throughput. This study answers the research questions (1) What problems do flow simulations identify when analyzing the performance and bottlenecks of packaging lines in the pharmaceutical industry? (2) What effects does the proposed solution have on the performance of the bottleneck flow in the production process? By simulating different scenarios and identifying problems related to resource usage and priorities at join points, solutions to bottlenecks can be identified. By smoothing priorities and using buffer stocks, companies can create a smoother flow and improve the resilience of their manufacturing systems. The study demonstrates that design and layout choices have a significant effect on total throughput, which emphasizes the importance of careful layout planning and optimization within the pharmaceutical industry. By improving key metrics such as mean time between failure and mean time to repair, companies can increase throughput, productivity and efficiency while achieving financial, ecological, and social sustainability. Recommendation for future studies includes studying the effect of operator location and machine speed on waiting time.

The Scandinavian brown bear went through a severe bottleneck due to human actions in the early 1900’s that reduced the population to low numbers. After protective measures were taken, the population recovered and today it consists of around 3000 individuals. Such loss of genetic variation can have long-lasting effects on population viability even as populations recover, and is therefore important to consider in the management and conservation of species. Previous studies investigating the genetic effects of the bottleneck based on genetic markers, such as microsatellites and mitochondrial DNA, have rendered inconclusive results. Here, whole-genome sequencing of historical and contemporary bears was used to estimate heterozygosity, inbreeding and mutational load in the pre- and post- bottleneck population. Surprisingly, a significant increase in heterozygosity was found in the contemporary population and no significant increase in inbreeding over time was found. However, mutational load was higher in the contemporary subpopulation in the south of Sweden compared to the subpopulation in the north, and the southern subpopulation seems to have lost unique genetic variation after the bottleneck. The results indicate that although the population was negatively affected by the bottleneck, a following increase in gene flow could have contributed to successful recovery of the population.

Detta examensarbete är utfört på Scania Södertälje i syftet att kartlägga cykeltider av Kärnamakeri avdelning på nya gjuteriet. Mätning alla av cykeltider för att sen kartläggas görs för att tydligt visualisera utrustnings samt avdelnings kapacitet. Projektet inleddes med insamling av all nödvändig information som underlag för genomförande av arbetet och därefter valdes lämpliga metoder för att uppnå målen. Resultatet visar en berglipig cykeltidskartläggning över hela avdelningen med en tydlig identifiering av flaskhalsarna som styr hela produktionen. Med hjälp av den valda metoden analyserades flaskhalsarna djupnade för att hitta möjliga slöserier i form av onödiga rörelser för att utföra processen. Förbättringsförslag presenterades som avslutning del av projektens resultat samt även som rekommendationer för fortsatt förbättringsarbete att optimera produktionskapacitet. / This thesis project is implemented at Scania Södertälje with the aim of mapping cycle times of the core making department at the new foundry. All measurement of cycle times to then be mapped is done to clearly visualize equipment and department capacity. The project began with the collection of all necessary information as a basis for carrying out the work and then suitable methods were chosen to achieve the goals. The result shows a pronounced cycle time mapping across the entire department with a clear identification of the bottlenecks that control the entire production. Using the chosen method, the bottlenecks were analysed in depth to find possible waste in the form of unnecessary movements to carry out the process. Improvement proposals were presented as a concluding part of the project's results and also as recommendations for continued improvement work to optimize production capacity.

No description available.

No description available.

No description available.

Norrmejerier’s supply of raw material fluctuates during different periods of the year. During the months where the company receives the most milk there is a need of an increase in cheese production in order to refine the milk in a more value creating way. To enable an increase in the production the process of the cheese department in Umeå is mapped. To determine the critical parts of the process a month’s production is simulated in a discrete event system.  Today the department produces eight different cheese types of which they aim to carry a constant partition. All process times except the one for the last process step are independent on the cheese type produced. The process time for the last leg differs with up to 70% depending on the cheese type. Hence the total production volume depends on the order in which the different types of cheese are produced, the production sequence.  In this thesis we optimize the production sequence with a genetic search algorithm. Given the predetermined distribution of cheese types a set of production sequences is created which are tested against a model imitating the process. The set of sequences is gradually improved through the genetic algorithm to finally reach a solution considered good enough. This enables us to derive the maximum capacity of the process and determine its bottlenecks.  This study concludes that a production increase of 16,6% can be achieved by additional staff, adjusted work routines and an optimized production sequence. To further increase the production, an increased capacity of the first part of the process is required, that being the pasteurization step. Given an increase in the pasteurization step the next bottleneck would appear in the last part of the process, where the cheese is placed in a salt brine. / Norrmejerier har under året en fluktuerande invägning av råvara som vid toppar ger upphov till ett råvaruöverskott. För att förädla råvaran på ett mer värdeskapande vis finns ett behov av en kapacitetsökning av osttillverkning under de perioder då invägningen är som störst. För att möjliggöra en ökad produktion av ost kartläggs processen på ysteriavdelningen i Umeå. Avdelningen producerar åtta osttyper och önskar bibehålla en förutbestämd fördelning av dessa. Givet fördelningen simuleras en månads produktion i ett diskret händelsesystem där kritiska delar i processen identifieras.  Beläggningstiden för processdelarna på ysteriavdelningen är mer eller mindre oberoende av vilken osttyp som produceras med undantag för det sista processteget där beläggningstiden varierar med upp till 70%. Detta leder till att den totala produktionsvolymen är beroende av i vilken ordning de olika osttyperna produceras.  I denna studie optimerar vi denna ordningsföljd, benämnd produktionssekvens, med en genetisk sökalgoritm. Givet den förutbestämda fördelningen av osttyper skapar vi en mängd produktionssekvenser som testas mot en modell som imiterar processen. Mängden produktionssekvenser utvecklas successivt via den genetiska algoritmen för att kulminera i en tillräckligt bra lösning. Genom att göra detta kan processens flaskhals identifieras och en maxkapacitet härledas.  Studien fastställer att förändringar i bemanning, arbetssätt och en optimerad produk- tionssekvens kan ge en ökad produktion om 16,6%. För ytterligare förbättringar krävs ökad kapacitet på den första processdelen där mjölken pastöriseras. Om kapaciteten utökas uppstår nästa flaskhals i det sista processteget, saltlaken.

This simulation study was carried out at a company in the automotive industry. In their factory, the company manufactures gearboxes, where the gearbox assembly is one step in the manufacturing process. In the main assembly line, many different gearbox variants are manufactured, each with its own cycle time in the different assembly stations. Due to this variation in cycle times the throughput in the main assembly line can fluctuate depending on which gearbox variants are produced. To maintain a desired throughput a tool for forecasting the throughput would be of great value to the company. Due to the large number of gearbox variants, the company also experiences difficulties understanding if there are stations in the assembly which are recurring bottlenecks. The purpose of this project is to create a tool that can forecast the daily throughput of the main assembly line and identify bottlenecks in the main assembly line. The purpose is also to investigate whether some stations often are bottleneck stations during a certain period. The model evaluation was based on validation of input-output-transformation. A bottleneck analysis based on the active times of the stations was carried out with the best-performing model. The best-performing model version resulted in 68 percent valid forecasts for the 57 historical production days. The forecasts were valid for all historical production days which were Mondays, Tuesdays, Wednesdays, and Thursdays. For Fridays and Sundays, 25 and 0 percent were valid, respectively. A bottleneck analysis was carried out for 37 historical production days where the result showed that, based on the share of active time, a single station was identified as the bottleneck station for all days. The difference between the bottleneck station’s share of active time and other stations’ share of active time was deemed small enough to be within the error margin. Additionally, the queues in the main assembly line indicated that the bottleneck was located within another group of stations. Due to the ambiguity of the results, the conclusion drawn was that the main assembly line did not experience any clear bottleneck station during the 37 historical production days. / <p>Examensarbetet är utfört vid Institutionen för teknik och naturvetenskap (ITN) vid Tekniska fakulteten, Linköpings universitet</p>