This research evaluates data-driven demand models using natural language processing techniques and analytical approaches. The first essay offers a comprehensive review of data-driven newsvendor literature and applies natural language processing techniques, including latent semantic analysis, latent Dirichlet allocation and cluster analysis to analyze the text data. This study highlights emerging trends and future research directions in the field of data-driven newsvendor research. The second essay contributes to the data-driven newsvendor inventory management literature by proposing nonparametric approaches that include Tobit and quantile regression incorporating leverage values under conditions of homogeneity and heterogeneity. Lastly, the third essay addresses the optimization of healthcare facility location and resource allocation in post-earthquake scenarios, presenting a linear programming model with telemedicine integration for effective disaster response. This study applies the model to the 2005 Kashmir earthquake in Pakistan. These essays collectively highlight the potential of data-driven methodologies in enhancing decision-making processes across diverse domains, while also pointing towards future research directions to address inherent complexities and uncertainties of the models.

No / We present a new skin deformation method to create dynamic skin deformations in this paper. The core elements of our approach are a dynamic deformation model, an efficient data-driven finite difference solution, and a curve-based representation of 3D models.We first reconstruct skin deformation models at different poses from the taken photos of a male human arm movement to achieve real deformed skin shapes. Then, we extract curves from these reconstructed skin deformation models. A new dynamic deformation model is proposed to describe physics of dynamic curve deformations, and its finite difference solution is developed to determine shape changes of the extracted curves. In order to improve visual realism of skin deformations, we employ data-driven methods and introduce skin shapes at the initial and final poses in to our proposed dynamic deformation model. Experimental examples and comparisons made in this paper indicate that our proposed dynamic skin deformation technique can create realistic deformed skin shapes efficiently with a small data size.

No / We introduce a new "marker-based" model for use in capturing equine movement. This model is informed by a sound biomechanical study of the animal and can be deployed in the pursuit of many undertakings. Unlike many other approaches, our method provides a high level of automation and hides the intricate biomechanical knowledge required to produce realistic results. Due to this approach, it is possible to acquire solved data with minimal manual intervention even in real-time conditions. The approach introduced can be replicated for the production of many other animals. The model is first informed by the veterinary world through studies of the subject's anatomy. Second, further medical studies aimed at understanding and addressing surface processes, inform model creation. The latter studies address items such as skin sliding. If not otherwise corrected these processes may hinder marker based capture. The resultant model has been tested in feasibility studies for practicality and subject acceptance during production. Data is provided for scrutiny along with the subject digitally captured through a variety of methods. The digital subject in mesh form as well as the motion capture model aid in comparison and show the level of accurateness achieved. The video reference and digital renders provide an insight into the level of realism achieved.

<p dir="ltr">Increased frequency and intensity of disasters necessitate the dynamic post-disaster recovery process. Developing human mobility patterns, household return decision-making models, and agent-based simulations in disaster management has opened a new door towards more intricate and enduring recovery frameworks. Despite these opportunities, the importance of a unified framework is underestimated to identify the underlying mechanisms hindering the post-disaster recovery process. My research has been geared towards forging advancements in civil and disaster management, focusing on two main areas: (1) modeling the post-disaster recovery process and (2) identifying critical transitions within the recovery process.</p><p dir="ltr">My dissertation explores the collective and individual dynamics of post-disaster recovery across different spatial and temporal scales. I have identified the best recovery strategies for various contexts by constructing data-driven socio-physical multi-agent systems. Employing various advanced computational methodologies, including machine learning, system dynamics, causal discovery, econometrics, and network analysis, has been instrumental. I start with aggregated level analysis for post-disaster recovery. Initially, I examined the system dynamics model for the post-discovery recovery process in socio-physical systems, using normalized visit density of points of interest and power outage information. Through counterfactual analyses of budget allocation strategies, I discovered their significant impact on recovery trajectories, noting that specific budget allocations substantially enhance recovery patterns. I also revealed the urban-rural dissimilarity by the data-driven causal discovery approach. I utilized county-level normalized visit density of points of interest and nighttime light data to identify the relationship between counties. I found that urban and rural areas have similar but different recovery patterns across different types of points of interest.</p><p dir="ltr">Moving from aggregated to disaggregated level analysis on post-disaster recovery, I investigated household-level decision-making regarding disaster-induced evacuation and return behaviors. The model yielded insights into the varying influences of certain variables across urban and rural contexts. Subsequently, I developed a unified framework integrating aggregated and disaggregated level analyses through multilayer multi-agent systems to model significant shifts in the post-disaster recovery process. I evaluated various scenarios to pinpoint conditions for boosting recovery and assessing the effects of different intervention strategies on these transitions. Lastly, a comparison between mathematical models and graph convolutional networks was conducted to better understand the conditions leading to critical transitions in the recovery process. The insights and methodologies presented in this dissertation contribute to the broader understanding of the disaster recovery process in complex urban systems, advocating for a shift towards a unified framework over individual models. By harnessing big data and complex systems modeling, I can achieve a detailed quantitative analysis of the disaster recovery process, including critical transition conditions of the post-disaster recovery. This approach facilitates the evaluation of such recovery policies through inter-regional comparisons and the testing of various policy interventions in counterfactual scenarios.</p>

京都大学 / 新制・課程博士 / 博士(工学) / 甲第24146号 / 工博第5033号 / 新制||工||1786(附属図書館) / 京都大学大学院工学研究科航空宇宙工学専攻 / (主査)教授 藤本 健治, 教授 松野 文俊, 教授 森本 淳 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

In many industrial and scientific applications, accurate sensor measurements are crucial. Instruments such as nitrate sensors are vulnerable to environmental conditions, calibration drift, high maintenance costs, and degrading. Researchers have turned to advanced computational methods, including mathematical modeling, statistical analysis, and machine learning, to overcome these limitations. Deep learning techniques have shown promise in outperforming traditional methods in many applications by achieving higher accuracy, but they are often criticized as 'black-box' models due to their lack of transparency. This thesis presents a framework for deep learning-based soft sensors that can quantify the robustness of soft sensors by estimating predictive uncertainty and evaluating performance across various scenarios. The framework facilitates comparisons between hard and soft sensors. To validate the framework, I conduct experiments using data generated by AI and Cyber for Water and Ag (ACWA), a cyber-physical system water-controlled environment testbed. Afterwards, the framework is tested on real-world environment data from Alexandria Renew Enterprise (AlexRenew), establishing its applicability and effectiveness in practical settings. / Master of Science / Sensors are essential in various industrial systems and offer numerous advantages. Essential to measurement science and technology, it allows reliable high-resolution low-cost measurement and impacts areas such as environmental monitoring, medical applications and security. The importance of sensors extends to Internet of Things (IoT) and large-scale data analytics fields. In these areas, sensors are vital to the generation of data that is used in industries such as health care, transportation and surveillance. Big Data analytics processes this data for a variety of purposes, including health management and disease prediction, demonstrating the growing importance of sensors in data-driven decision making. In many industrial and scientific applications, precision and trustworthiness in measurements are crucial for informed decision-making and maintaining high-quality processes. Instruments such as nitrate sensors are particularly susceptible to environmental conditions, calibration drift, high maintenance costs, and a tendency to become less reliable over time due to aging. The lifespan of these instruments can be as short as two weeks, posing significant challenges. To overcome these limitations, researchers have turned to advanced computational methods, including mathematical modeling, statistical analysis, and machine learning. Traditional methods have had some success, but they often struggle to fully capture the complex dynamics of natural environments. This has led to increased interest in more sophisticated approaches, such as deep learning techniques. Deep learning-based soft sensors have shown promise in outperforming traditional methods in many applications by achieving higher accuracy. However, they are often criticized as "black-box" models due to their lack of transparency. This raises questions about their reliability and trustworthiness, making it critical to assess these aspects. This thesis presents a comprehensive framework for deep learning-based soft sensors. The framework will quantify the robustness of soft sensors by estimating predictive uncertainty and evaluating performance across a range of contextual scenarios, such as weather conditions, flood events, and water parameters. These evaluations will help define the trustworthiness of the soft sensor and facilitate comparisons between hard and soft sensors. To validate the framework, we will conduct experiments using data generated by ACWA, a cyber-physical system water-controlled environment testbed we developed. This will provide a controlled environment to test and refine our framework. Subsequently, we will test the framework on real-world environment data from AlexRenew. This will further establish its applicability and effectiveness in practical settings, providing a robust and reliable tool for sensor data analysis and prediction. Ultimately, this work aims to contribute to the broader field of sensor technology, enhancing our ability to make informed decisions based on reliable and accurate sensor data.

In this thesis, the focus is on enhancing EMI's products, Onboard and PocketMine leading software solutions in the mining sector. This study explores how the extensive data gathered by Onboard can be used to develop a more precise and reliable positioning system for PocketMine and to create the foundation of redundancy for Onboard using machine learning models. Furthermore, it will explore what machine learning model performs optimally with this data. This thesis is motivated by the potential of data-driven methodologies to enhance the safety and accuracy of EMI’s products, significantly improving operational safety and precision in challenging underground environments but also contributing to the broader field of positioning technology. The goals for this thesis are achieved by comparing four different ML models on three distinct datasets based on locations in the mine to decide which models the final solution will be using. Additionally, the idea of creating a model encapsulating the entire mine is examined and compared to the POI-specific models to see if it is feasible for one model to learn the intricacies of the mine. In addition to this, the deployment strategy will be discussed. Upon comparing the models against each other and the mine-wide model, it was decided to move on with Weighted K Nearest Neighbors as the model of choice based on several evaluation metrics. The large scale of the mine proved too great to be handled by one model so the decision to cluster the mine into 100 distinct clusters and create one model for each cluster was made.  The results show that the proposed solution made a great improvement in positional accuracy over the current positioning algorithm of PocketMine. This improvement suggests in line with testing it against Onboard that the proposed model could effectively serve as a reliable backup system for Onboard.

Purpose - This study explores how knowledge exchange between sales and marketing can enhance data-driven marketing initiatives for firms in the high-tech process industry. Additionally, the study aims to identify the factors that drive alignment and collaboration between sales and marketing interfaces from a dynamic capability perspective. Method - This master's thesis is an exploratory study with an inductive approach. 10 qualitative interviews were conducted with employees from a high-tech process automation company, specifically those working in marketing and sales roles. The interviews follow a semi-structured approach, and a thematic analysis was performed to examine the empirical findings. Findings - The study emphasizes the significance of collaboration, knowledge exchange, and functional alignment between sales and marketing, in the context of data-driven marketing and the sales lead generation. By applying a dynamic capability framework, the study sheds light on how firms can leverage knowledge exchange and functional alignment to capitalize on market opportunities and gain a competitive advantage.  Theoretical and practical contributions - The study delves into the underexplored realm of data-driven marketing within the high-tech process industry, particularly focusing on the intricate dynamics between sales and marketing functions during the lead generation process. Through its analysis, the research not only enriches theoretical understanding but also offers practical insights for managers in the high-tech process industry, providing some recommendations to enhance collaboration, knowledge exchange, and optimize data-driven marketing initiatives. / Syfte - Denna studie undersöker hur kunskapsutbyte mellan försäljning och marknadsföring kan förbättra datadrivna marknadsföring initiativ för företag inom högteknologisk processindustri. Dessutom syftar studien till att identifiera de faktorer som driver linjering och samarbete mellan försäljning och marknadsföring, ur ett perspektiv från teorier inom Dynamic Capabilities.  Metod - Denna uppsats är en explorativ studie med en induktiv ansats. 10 kvalitativa intervjuer har genomförts med anställda på ett företag inom högteknologisk process automation, specifikt de som arbetar inom marknadsföring och försäljning. Intervjuerna har följt en semi-strukturerad metod och en tematisk analys har genomförts för att undersöka de empiriska resultaten. Resultat - Studien betonar vikten av samarbete, kunskapsutbyte och funktionell linjeringen mellan försäljning och marknadsföring, i kontexten av datadriven marknadsföring och sales lead generation. I studien har teorier inom Dynamic Capabilities belyst studien, mer specifikt hur företag kan utnyttja kunskapsutbyte och funktionell linjeringen för att utnyttja marknadsmöjligheter och skapa konkurrensfördelar. Teoretiska och praktiska bidrag - Studien utforskar det underutvecklade området datadriven marknadsföring inom högteknologisk processindustri, med särskilt fokus på de komplexa dynamikerna mellan försäljnings- och marknadsförings funktioner under processen för sales lead generation. Genom sin analys berikar forskningen inte bara den teoretiska utan erbjuder också praktiska insikter för chefer inom högteknologisk processindustri, med rekommendationer för att förbättra samarbete, kunskapsutbyte och optimera datadrivna marknadsföring initiativ.

Artificial intelligence is a technology that many individuals view favourably intoday's world. Where the technology contributes to a variety of benefits when usedthat can help individuals perform specific tasks. It has been shown from previousresearch that individuals' willingness to use artificial intelligence is affected by thetrust they have in the technology. It discusses how working methods and processesare affected by artificial intelligence and how it is of utmost importance to promotethe individual's trust in the technology to later promote the way artificialintelligence is used. In the previous research, various methods and models havebeen identified and demonstrated to measure and build appropriate trust in AI aswell as for various variables that affect the individual's trust and thus willingnessto use artificial intelligence as a technology.The study aims to study trust in and use of artificial intelligence from a user'sperspective where the individual is in focus. The study thus aims to study theimpact of trust on the use of the technology, any variables that affect users' trustin the technology and what is important in the implementation and use of artificialintelligence to promote trust and the use of artificial intelligence. This is tocontribute with nuanced knowledge of what is required to promote effectivecollaboration between humans and artificial intelligence when performing varioustasks.The study contains a qualitative and inductive approach where empirical data wascollected through open individual interviews and observations where respondentsinteracted with an AI tool and where they performed simple use cases. Throughcontent analysis, the empirical material has created a nuanced andknowledgeexpanding view of the phenomenon and identified aspects andvariables with an impact on individuals' trust and use of artificial intelligence.These are areas that, according to respondents, have a major impact on decisionsto use or not use artificial intelligence tools.The results of the study show that there are aspects and variables around trust thathave a major impact on decisions to use AI tools. These are thus important to keepin mind when implementing AI tools to promote trust and thus the use of artificialintelligence

Computational fluid dynamics (CFD) has been widely used to simulate turbulent flows. Although an increased availability of computational resources has enabled high-fidelity simulations (e.g. large eddy simulation and direct numerical simulation) of turbulent flows, the Reynolds-Averaged Navier-Stokes (RANS) equations based models are still the dominant tools for industrial applications. However, the predictive capability of RANS models is limited by potential inaccuracies driven by hypotheses in the Reynolds stress closure. With the ever-increasing use of RANS simulations in mission-critical applications, the estimation and reduction of model-form uncertainties in RANS models have attracted attention in the turbulence modeling community. In this work, I focus on estimating uncertainties stemming from the RANS turbulence closure and calibrating discrepancies in the modeled Reynolds stresses to improve the predictive capability of RANS models. Both on-line and off-line data are utilized to achieve this goal. The main contributions of this dissertation can be summarized as follows: First, a physics-based, data-driven Bayesian framework is developed for estimating and reducing model-form uncertainties in RANS simulations. An iterative ensemble Kalman method is employed to assimilate sparse on-line measurement data and empirical prior knowledge for a full-field inversion. The merits of incorporating prior knowledge and physical constraints in calibrating RANS model discrepancies are demonstrated and discussed. Second, a random matrix theoretic framework is proposed for estimating model-form uncertainties in RANS simulations. Maximum entropy principle is employed to identify the probability distribution that satisfies given constraints but without introducing artificial information. Objective prior perturbations of RANS-predicted Reynolds stresses in physical projections are provided based on comparisons between physics-based and random matrix theoretic approaches. Finally, a physics-informed, machine learning framework towards predictive RANS turbulence modeling is proposed. The functional forms of model discrepancies with respect to mean flow features are extracted from the off-line database of closely related flows based on machine learning algorithms. The RANS-modeled Reynolds stresses of prediction flows can be significantly improved by the trained discrepancy function, which is an important step towards the predictive turbulence modeling. / Ph. D. / Turbulence modeling is a critical component in computational fluid dynamics (CFD) simulations of industrial flows. Despite the significant growth in computational resources over the past two decades, the time-resolved high-fidelity simulations (e.g., large eddy simulation and direct numerical simulation) are not feasible for engineering applications. Therefore, the small-scale turbulent velocity fluctuations have to resort to the time-averaging modeling. Reynolds-averaged Navier-Stokes (RANS) equations based turbulence models describe the averaged flow quantities for turbulent flows and are believed to be the dominant tools for industrial applications in coming decades. However, for many practical flows, the predictive accuracy of RANS models is largely limited by the model-form uncertainties stemming from the potential inaccuracies in the Reynolds stress closure. As RANS models are used in the design and safety evaluation of many mission-critical systems, such as airplanes and nuclear power plants, properly estimating and reducing these model uncertainties are of significant importance. In this work, I focus on estimating uncertainties stemming from the RANS turbulence closure and calibrating discrepancies in the modeled Reynolds stresses to improve the predictive capability of RANS models. Several data-driven approaches based on stateof-the-art data assimilation and machine learning algorithms are proposed to achieve this goal by leveraging the use of on-line and off-line high-fidelity data. Numerical simulations of several canonical flows are used to demonstrate the merits of the proposed approaches. Moreover, the proposed methods also have implications in many fields in which the governing equations are well understood, but the model uncertainties come from unresolved physical processes.