C Mathematical and Quantitative Methods
Refine
Document Type
- Bachelor Thesis (6)
- Master's Thesis (6)
- Article (1)
- Study Thesis (1)
Has Fulltext
- yes (14)
Keywords
- Lebensversicherung (2)
- Rettungswesen (2)
- Risikomanagement (2)
- Versicherungsmathematik (2)
- API (1)
- Ameisenalgorithmus (1)
- Bewertung (1)
- Bewertungsmethode (1)
- Bilanzierung (1)
- Blue Carbon Potential Index (1)
Faculty
- Fakultät 10 / Institut für Informatik (5)
- Fakultät 09 / Institut für Rettungsingenieurwesen und Gefahrenabwehr (4)
- Fakultät 04 / Institut für Versicherungswesen (2)
- Fakultät 12 / Institut für Technologie und Ressourcenmanagement in den Tropen und Subtropen (2)
- Fakultät 07 / Institut für Nachrichtentechnik (1)
A test tool for Langton's ant-based algorithms is created. Among other things, it can create test files for the NIST-Statistical-Test-Suite. The test tool is used to investigate the invertibility, ring formation and randomness of 7 created models which are extensions of Langton’s ant. The models are examined to possibly use them as pseudo-random generator (PRG) or block cipher. All models use memories which are based on tori. This property is central, because this is how rings are formed in the first place and in addition the behavior of all models at the physical boundaries of the memory is clearly defined in this way. The different models have special properties which are also investigated. These include variable color sets, discrete convolution, multidimensionality, and the use of multiple ants, which are arranged fractal hierarchically and influence each other. The extensions convolution, multidimensional scalable and multidimensional scalable fractal ant colony are presented here for the first time. It is shown that well-chosen color sets and high-dimensional tori are particularly well suited as a basis for Langton's ant based PRGs. In addition, it is shown that a block cipher can be generated on this basis.
As a customer, it can be frustrating to face an empty shelf in a store. The market does not always realize that a product has been out of stock for a while, as the item is still listed as in stock in the inventory management system. To address this issue, a camera should be used to check for Out-of-Stock (OOS) situations.
This master thesis evaluates different model configurations of Artificial Neural Networks (ANNs) to determine which one best detects OOS situations in the market using images. To create a dataset, 2,712 photos were taken in six stores. The photos clearly show whether there is a gap on the shelf or if the product is in stock. Based on the pre-trained VGG16 model from Keras, two fully connected layers were implemented, with 36 different ANNs differing in the optimization method and activation function pairings. In total, 216 models were generated in this thesis to investigate the effects of three different optimization methods combined with twelve different activation function pairings. An almost balanced ratio of OOS and in-stock data was used to generate these models.
The evaluation of the generated OOS models shows that the FTRL optimization method achieved the least favorable results and is therefore not suitable for this application. Model configurations using the Adam or SGD optimization methods achieve much better results. Of the top six model configurations, five use the Adam optimization method and one uses SGD. They all achieved an accuracy of at least 93% and were able to predict the Recall for the OOS class with at least 91%.
As the data ratio between OOS and in-stock data did not correspond to reality in the previously generated models, the in-stock images were augmented. Including the augmented images, new OOS models were generated for the top six model configurations. The results of these OOS models show no convergences. This suggests that more epochs in the training phase lead to better results. However, the results of the OOS model using the Adam optimization method and the Sigmoid and ReLU activation functions stand out positively. It achieved the best result with an accuracy of 97.91% and a Recall of the OOS class of 87.82%.
Overall, several OOS models have the potential to increase both market sales and customer satisfaction. In a future study, the OOS models should be installed in the market to evaluate their performance under real conditions. The resulting insights can be used for continuous optimization of the model.
Mangrove forests have been studied broadly in the recent three decades for their outstanding ability to sequester carbon in the beneath soil and other beneficial ecosystem services. Endeavors to conserve and regenerate mangrove cover are still increasing worldwide as a mechanism to include them in NDCs and carbon markets. Therefore, decision-makers in the private and public sectors require identify possible areas for conservation and restoration prior to blue carbon project investment. Thus, an integral assessment of potential mangrove carbon reservoirs in a landscape scale, considering environmental and socioeconomic factors was performed. This study was aimed to determine areas with the highest blue carbon sequestration potential in the Gulf of Guayaquil through the construction of a Blue Carbon Potential Index (BCPI) based on Spatial Multicriteria Analysis (SMCA). A narrative integrative literature review was employed to select indicators of mangrove carbon sequestration gains and losses. These indicators were pondered following the Analytical Hierarchy Process (AHP) with the judgments of two experts and reclassified in four potential categories based on their thresholds. Since no consensus was achieved in the indicator importance hierarchization, a comparative of equal weighting method and AHP weighting was implemented. The linear combination rule was used to integrate these factors into a unique-scaled index supported by a geographic Information System (GIS). The results showed that 15.82% and 16.21% of the study area belonged to high and moderate potential of blue carbon sequestration respectively. Moreover, no significant differences were found between the two weighting methods applied. The BCPI provides a comprehensive understanding of spatial distribution of blue carbon potential reservoirs and grants a quantification of this potential to prioritize conservation and restoration areas.
Diese Arbeit wertet Leistellendaten von zwei Landkreisen aus und untersucht dabei, welche Veränderung hinsichtlich der mittleren Anzahl und Dauer von Rettungsdiensteinsätzen im Zeitraum der COVID-19-Pandemie aufgeteilt nach Stunden- und Tageskategorien besteht. Anschließend werden die Veränderungen von Anzahl und Dauer beider Landkreise verglichen, um diese auf Unterschiede zu prüfen. Da in der aktuellen Literatur unzureichend dargelegt ist, wie Veränderungen des Einsatzanzahl oder -dauer im Rettungsdienst zwischen Kreisen sowie über die Tage und Tagesstunden verteilt sind, trägt diese Untersuchung dazu bei diese Forschungslücke zu schließen. Die statistische Auswertung der Leitstellendaten umfasst mehrere Schritte, sodass zuerst eine Aufbereitung durchgeführt wird, bevor die mittlere Einsatzanzahl und -dauer für vier Zeiträume sowie die prozentuale Veränderung zwischen einem Zeitraum vor und drei Zeiträumen während der Pandemie ermittelt werden. Im Anschluss wird die Veränderung mittels Kruskal-Wallis-Test beziehungsweise Varianzanalyse nach Welch auf Signifikanz überprüft. Abschließend erfolgt ein qualitativer Vergleich zwischen den Kreisen. Die Untersuchung zeigt, dass hinsichtlich Einsatzanzahl und -dauer zu allen drei Zeitraumen in der Pandemie Zu- und Abnahmen auftreten. Die Anzahl nimmt zum ersten Zeitraum in beiden Kreisen überwiegend ab, sowie zum vierten Zeitraum mehrheitlich zu, während zum zweiten Zeitraum im Lahn-Dill-Kreis eine häufigere Abnahme und im Main-Taunus-Kreis eine geringfügig überwiegende Zunahme vorliegt. Die Einsatzdauer unterliegt zum zweiten Zeitraum im Lahn-Dill-Kreis einer überwiegenden Zunahme und im Main-Taunus-Kreis einer überwiegenden Abnahme, während zum dritten sowie vierten Zeitraum in beiden Kreisen einer Steigerung stattfindet. Im Vergleich beider Gebietskörperschaften zeigen bei der Einsatzanzahl die Veränderungen zum zweiten und vierten Zeitraum und bei der Dauer die Veränderungen zum dritten und vierten Zeitraum mit der Mehrheit der Stundenkategorien Gemeinsamkeiten. Das Pandemiegeschehens sowie die nichtpharmazeutischen Interventionen sowie deren Folgen stellen Faktoren für die Veränderung von Einsatzanzahl und -dauer dar. Die geografische Lage sowie die lokale Ausgestaltung von Rettungsdienst und Krankentransport begründen, zusätzlich zu den oben genannten Aspekten, Unterschiede zwischen den Kreisen.