Doctor of Philosophy, The Ohio State University, 2024, Industrial and Systems Engineering

Sociotechnical systems in safety-critical domains are distributed and contain interdependencies between the different elements, including human and automated roles that need to coordinate and synchronize their activities with dynamic events in the environment. The advancement of technology and the introduction of machines capable of acting at a higher level of autonomy has increased the complexity of such Distributed Work Systems (DWSs). An envisioned DWS is described by a set of static paper-based documents and will be deployed in the next few years. The short-range low-altitude air mobility system is one very good example of an envisioned DWS. Interactions between human and automated roles and their environment are dynamic, evolve, and change over time, causing emergent effects like taskload peaks and coordination breakdowns. A well-designed DWS will be able to keep pace with the work environment dynamics (like the dynamics of aircraft governed by laws of flight in a short-range low-altitude air mobility system) and succeed in responding to the disturbance. This creates the need to understand the dynamics of envisioned DWS, such as how a DWS performs in high-paced situations like anomaly response. Assessing the feasibility and robustness of an envisioned DWS comes with challenges: the physical system does not yet exist, its design and operations are often underspecified, and multiple versions may exist within a designer community about what future operations will look like. Therefore, as a part of this dissertation, an exploratory early-stage computational modeling and simulation technique is described and demonstrated to evaluate an envisioned DWS. Using functional modeling and computational simulation capabilities, the dissertation shows a technique that can help evaluate envisioned DWS by discovering things that are not uncovered by traditional normative simulations. The primary advantage of the technique is the ability to evaluate the dynamics of work in (open full item for complete abstract)

Doctor of Philosophy, The Ohio State University, 2024, Industrial and Systems Engineering

Additive manufacturing overcomes limitations of conventional manufacturing ranging from geometric complexity to multi-material integration. This newer technology requires studies into how the process affects the structure and material properties of the finished part. Traditionally, this is done through experimental changing of process parameters and destructive testing of the manufactured material. This traditional process is time consuming and cost ineffective. One solution for this issue is a data driven approach for digitizing the manufacturing exploration process. This thesis introduces a framework for investigating the relationships between additive manufacturing process parameters and observed structures. A characterization method is presented followed by separate microstructure and mesostructure modeling methods. A traditional feature statistics approach is applied to the microstructure while a machine learning computer vision approach is applied to the mesostructure. The combination of these two approaches creates a workflow for fully understanding how the manufacturing process parameters affect the varying levels of material structure.

Doctor of Philosophy (PhD), Ohio University, 2024, Industrial and Systems Engineering (Engineering and Technology)

The healthcare supply chain is a network of physical and informational resources required to deliver a good or service to patients. The vaccine supply chain is a type of healthcare supply chain that starts with raw materials. After going through the manufacturing process, the vaccines are distributed to warehouses and then to health providers, hospitals, and pharmacies. Finally, the vaccines are given to patients. Like many other supply chain problems, minimizing the expenditure and improving the supply chain network is crucial. This study aims to develop a mathematical model for optimizing the vaccine supply chain network. A multi-echelon, multi-period location-allocation model considering inventory management and transshipment is developed to design and manage a vaccine supply chain network in order to minimize total cost. In this study various costs including location cost, transportation and transshipment costs, inventory holding costs, wastage cost and penalty cost for unsatisfied demands are considered. Transshipment—transporting products between facilities in the same echelon of the supply chain—is utilized as a strategy to overcome the demand fluctuation and preventing the stockout. Since vaccines can be scarce, it is important to design an efficient network for vaccine distribution based on the equity. For this reason, when making the allocation decisions, priority is given to patients based on their characteristics and the ratio of unmet demand to the population. In addition, this research examines ways to improve inventory management by considering approaches such as transshipment between health providers. To make the model more realistic, a robust model is considered to deal with uncertainty, especially demand uncertainty. The models are solved by using CPLEX software to find the optimal solutions. Also, the sensitivity analysis is used to check the model's response to different scenarios and make sure that the models are giving answers as (open full item for complete abstract)

Master of Science, The Ohio State University, 2024, Industrial and Systems Engineering

The manufacturing process of injection molding is widely used around the world as a method to produce thermoplastic products. A challenge that comes with injection molding is selecting a combination of controllable process variables (CPVs) which will result in the production of an acceptable part. Process windows, a set of operating conditions that appear as a window-shaped area on a graph, provide a solution to this challenge by displaying which combination of CPVs will produce an acceptable injection molded part. Although a process window is useful tool, the procedure to construct one experimentally can be long, costly, and difficult. Injection molding simulation software presents an opportunity to develop process windows economically and effectively. Therefore, the goal is this thesis is to utilize simulation software to construct process windows. Before process windows were constructed, the injection molding simulation software was evaluated on its ability to predict the injection pressures of a simple model. The software's ability to predict the manufacturing results of a complex model were also evaluated. Once both evaluations were completed, process windows were constructed for both the filling and packing stages of the injection molding cycle. A methodology was developed for the construction of process windows allowing for modifications so that the process can be tailored to each processor's desired product.

Master of Science (MS), Ohio University, 2024, Industrial and Systems Engineering (Engineering and Technology)

In response to escalating inventory costs, dynamic purchasing needs, and the demand for rapid operations in the retail sector, both the warehousing and retail industries have accelerated their pace of innovation. Among these advances, the development of automated warehousing and storage systems stands out. However, despite widespread adoption, a comprehensive framework for effectively implementing these systems remains lacking. Hence, this study proposes a systematic approach that provides a foundational blueprint for harnessing vital information from historical sales data in the deployment of intelligent warehouse systems, incorporating a wide array of Automated Storage and Retrieval Systems (AS/RS) technologies. Specifically, it employs unsupervised machine learning for time series clustering to analyze historical sales data, while adapting and modifying the Recency, Frequency, Monetary (RFM) model to optimize the prioritized management of stock-keeping units (SKUs) in periodic segments.

Doctor of Philosophy, The Ohio State University, 2024, Industrial and Systems Engineering

Practitioners must synchronize to cope with the challenges of complex adaptive systems. Pre-planned synchronization will always be insufficient, forcing practitioners to re-synchronize in real time. This study analyzes cases from two work settings - healthcare and air traffic control - to reveal how practitioners modulate monitoring and their authority relationships as they navigate the goals, pressures, and tradeoffs of their various activities. This study revealed three durable patterns of monitoring modulation: rhythm, resolution, and extent. It also revealed five patterns of how people adapted authority relationships: take initiative, delegate, reassign, borrow, and prolong. In addition, it revealed fifteen signals that practitioners track and eight benefits of monitoring, as well as six key characteristics for authority relationships. These results highlight how practitioners adapt in real time to create a readiness to re-synchronize, despite limited organizational support.

Master of Science, The Ohio State University, 2024, Civil Engineering

During the rubber manufacturing process, a large amount of rubber waste is produced, which has traditionally been dumped into landfills. Despite regulations attempting to mitigate the impact of waste disposal, landfills continue to negatively affect soil, air, water, and natural life. A potential solution is to use waste rubber as a powder filler in place of cement and sand, providing a long-term solution to the environmental consequences of modern waste accumulation. This study investigates the effect of incorporating waste rubber into concrete to improve the environmental sustainability of concrete structures in the construction industry. Seven different rubber types were collected and ground into sizes comparable to cement and sand. Two of the rubber materials were un-vulcanized rubber, four were vulcanized rubber, and the last material was derived from discarded tires. Results showed that replacing cement with rubber resulted in a significant reduction in the compressive strength of mortars, attributed to the coarser particle size of the rubber compared to cement. Additionally, a slight reduction in the flowability of fresh mortar was observed. Concrete specimens were cast to investigate the effect of rubber as a sand replacement. Promising results were observed in terms of compressive strength: at 28-day, the range of compressive strength reduction was 2-12% for 5% replacement, 4-20% for 10% replacement, and 13-22% for 15% replacement. However, a significant reduction was observed in the modulus of elasticity, with an average reduction of 40% when 15% sand was replaced. The cost and environmental impact assessment indicated a reduction in CO2 emissions and the utilization of a significant amount of waste rubber, preventing it from being disposed of in landfills.

Doctor of Philosophy, The Ohio State University, 2024, Industrial and Systems Engineering

Manual materials handling work is a significant contributor to musculoskeletal injuries, particularly in the warehousing and storage sector. This dissertation aimed to study how people learn to use passive exoskeletons for manual materials handling tasks, similar to those performed in warehouses and distribution centers. The goal was to provide the data needed to allow workers and their organizations considering exoskeleton implementation to have reasonable expectations regarding the temporality of the learning process, and possibly provide insights regarding how to better temporally organize training sessions to facilitate the motor learning process and the adoption of passive exoskeletons. If adopted, various studies, including the present one, suggested that exoskeletons have the potential to reduce the injury rates in warehouse product selection jobs (Alemi et al., 2019; Baltrusch et al., 2020; Bosch et al., 2016; Qu et al., 2021; Wei et al., 2020). The current study assessed learning by measuring the changes in skill during retention and transfer conditions and comparing it to practice sets. The dependent variables for biomechanical performance were derived from the surface electromyographic (EMG) signals, trunk kinematic measures, and task durations. The current study had four main independent variables: exoskeleton conditions (with back-support passive exoskeleton vs. without), distributed practice methods, eight pick-to-placement height combinations, and four session types. Participants also reported their perceived level of effort, fatigue, and discomfort. In this research study, a total of 36 participants were recruited. The participants practiced in simulated picking tasks, which involved the transfer of twelve 10 kg boxes to the destination pallet and subsequently returning these boxes to the source pallet at four different height levels. The results showed that wearing the back-support passive exoskeleton significantly reduced muscle activity in (open full item for complete abstract)

Doctor of Philosophy, The Ohio State University, 2024, Industrial and Systems Engineering

Industrial automation has been rapidly improving in the past few decades due to advances in sensing, data storage, and artificial intelligence. However, human-robot interaction in these scenarios has been overlooked, leading to a current focus on human-centered automation. This approach can be beneficial for tasks such as small batch manufacturing where it is expensive to deploy automation due to lengthy manual reprogramming. How can an industrial robot learn a process plan directly from natural language? How can the user verify the correctness of the plan and observe what the robot is doing? It is crucial that the user can understand and verify the robot's actions especially in safety-critical manufacturing domains. It would be difficult to work with anyone, let alone a robot, if their behavior was unpredictable and they did not communicate. To this end, a system was designed that uses a Large Language Model (LLM) to convert natural language to sequences of robot actions. These sequences are displayed in a state machine to increase the observability and predicitability of the LLM output. Also, the state machine is executed in a virtual environment that matches the real-world environment. To test the performance of the system, user instructions were simulated and passed to the LLM along with problem context. 2 of the 5 prompts generated 12 of the 13 state machines that were both executable and met the goal in the user instructions. The LLM seemingly failed to plan in the other instances because the prompts did not link to pre-defined action names and the prompts described a goal and not actions to reach the goal. This indicates that LLMs are a potential avenue for reducing the programming burden by arranging robot actions directly from language. However, it does not seem appropriate to use them to directly generate plans given that this is an emergent capability and the LLM output changes significantly based on input prompts. A better appr (open full item for complete abstract)

Doctor of Philosophy, The Ohio State University, 2024, Industrial and Systems Engineering

To keep pace with growing globalization and technological change, reliance on third party entities by organizations is increasing. This requires agents to coordinate with other entities within the system, such as contractors and consultants, in ever increasing tangled layered networks and complex contexts. This process is not linear or additive but rather introduces new forms of work, surprise, and failure. Although there is a robust literature on how agent coordination, coordination across meaningful boundaries, such as, but not limited to, customer-contractor relationships, has not been subjected to comparable empirical research. This research examines cross-boundary joint activity in the construction and operations phases of the hyperscale data center industry, an industry characterized by exploding demand and routine use of third-party entities to build and operate these data centers for the hyperscale technology companies. The goal of this research is to understand how people coordinate with each other when they must coordinate across one or more boundaries and what constraints and asymmetries across boundaries are driving their actions. A corpus of cases was built using retrospective story elicitation. The corpus was analyzed using the Systemic Contributors and Adaptations Diagram (SCAD) method to identify the connection between pressures, conflicts, and adaptations that reveal aspects of cross-boundary joint activity. This analysis focused on identifying agent adaptations and the features of the network that created those interactions. Analysis of the cases found that projects critically rely on joint activity across roles that cross organizational boundaries and that this introduces constraints and asymmetries. Agents on all sides of the boundaries regularly engage in adaptations and adjustments to mitigate these constraints and asymmetries, effectively creating a covert work system required for project success. Patterns of adjustments and adaptations i (open full item for complete abstract)

Master of Science, Miami University, 2024, Mechanical and Manufacturing Engineering

Wire electrical discharge machining (WEDM) enables production of complex parts with tight tolerances, although maintaining dimensional accuracy in corners and tapers remains challenging due to wire deflection and vibration. This study optimizes WEDM parameters for achieving high-accuracy in machining complex geometrical parts and taper cuts in 6061 Aluminum alloy using Excetek W350G WEDM machine with a copper wire electrode. Parameters including Wire Tension, Pulse On-Time, Pulse Off-Time, Wire Feed Rate, Open Circuit Voltage, and Flashing Pressure were varied using L18 Taguchi Orthogonal Array and response graph method to identify optimal cutting conditions. Results indicated feature-specific optimization is crucial, as different geometrical features (rectangular fins, triangular fins, gears) exhibited varying critical parameters. Key findings highlighted the importance of Wire Tension and Pulse On-Time in maintaining cutting accuracy, although at varying levels for specific features. Response graphs demonstrated effects of major WEDM parameters on corner and profile accuracies, whereas Taguchi analysis provided optimum settings of parameters for each feature and taper cutting. Validation experiments for rectangular fins showed significant improvement in the dimensional error for the fin length and taper angle. These advancements will enhance precision, efficiency, and versatility of WEDM processes in machining complex profiles, and corners, contributing to precision manufacturing.

Doctor of Philosophy (PhD), Ohio University, 2024, Mechanical and Systems Engineering (Engineering and Technology)

In this research, we develop an innovative approach to assessing the reliability of complex engineering systems, which are typically characterized by multiple interdependent performance characteristics (PCs). Recognizing that the degradation of these PCs often follows a positive, increasing trend, we employ the gamma process as the foundational model for degradation due to its properties of independent and non-negative increments. A critical aspect of our model is the incorporation of random-effect bivariate Gamma process degradation models, which utilize a variety of copula functions. These functions are instrumental in accurately modeling the dependency structure between the PCs, a factor that significantly influences the overall system reliability. In conventional degradation modeling, fixed and predetermined failure thresholds are commonly used to determine system failure. However, this method can be inadequate as different systems may fail at varying times due to uncontrollable factors. Our model addresses this limitation by considering random failure thresholds, which enhances the accuracy of predicting when a system might fail. We implement a hierarchical Bayesian framework for the degradation modeling, data analysis, and reliability prediction processes. This approach is validated through the analysis of a practical dataset, demonstrating the model's applicability in real-world scenarios. Furthermore, our study responds to the increasing market demand for manufacturers to provide reliable information about the longevity of their products. Manufacturers are particularly interested in the 100p-th percentile of a product's lifetime distribution. Degradation tests are vital for this, as they offer insights into the product's lifespan under various conditions over time. Utilizing our proposed model, we propose a method for designing degradation tests. This method optimizes the number of systems to be tested, the (open full item for complete abstract)

Doctor of Philosophy, The Ohio State University, 2024, Industrial and Systems Engineering

First-order deterministic and stochastic optimization algorithms have gained significant importance in the past two decades primarily due to applications in data science and machine learning. In numerous recent problems, however, obtaining the exact gradient (in deterministic settings) or an unbiased gradient estimator (in stochastic settings) is computationally challenging. A naive implementation of the classical algorithms for such problems results in sub-optimal performance and unsatisfactory results. To address this gap, this research aims to investigate the properties of these problems, design new optimization algorithms, and investigate their theoretical convergence guarantees. Along this path, the dissertation makes three main contributions: 1) algorithmic development and in-depth analyses for a specific problem, namely stochastic composition optimization; 2) exploration of three biased stochastic approximation algorithms for the general setup and their theoretical analysis in the nonconvex setting; 3) investigation of accelerated gradient descent method for problems with inexact gradient oracles in convex setups and derivation of a new upper bound for the accumulated error. The first work focuses on the stochastic composition optimization problem. It explores scenarios where either the inner or outer function lacks Lipschitz continuity of their gradients. To generalize the assumption of Lipschitz continuity of the gradients, the notion of relative smoothness is introduced. The properties of composition gradients for three non-trivial combinations are examined, leading to a discussion of their corresponding smoothness properties. For each type of composition problem, first-order methods are proposed and their convergence analyses are conducted. Furthermore, the sample complexities for these proposed algorithms are established. The theoretical findings are then validated through experimental results. The subsequent research presents a unified fr (open full item for complete abstract)

Doctor of Philosophy, The Ohio State University, 2024, Industrial and Systems Engineering

Energy storage is widely used to respond to the uncertain balance of electricity supply and demand and prepare for the contingency. Among many purposes of energy storage, this dissertation will focus on arbitrage trade, peak load shift, and frequency regulation. For the first part, a two-stage stochastic programming model is introduced to schedule energy storage devices and maximize arbitrage profits for the storage operator. In addition, the model considers adjustments depending on the uncertain price of the real-time electricity market when the decision in the day-ahead market is made. Then, value of stochastic solution is computed to see effect of the stochastic programming. Furthermore, several interesting cases are observed and illustrated, such as simultaneous charging and discharging. These are considered as an sub-optimal solution in general, but this occurs in specific conditions. Second, when storage is used for peak load shift, it improves resource adequacy of the power systems by contribution of the power from energy storage. In this chapter, a non-performance penalty is imposed to ensure that energy storage operators reserve energy for such shortages. A stochastic dynamic programming model is used to obtain optimal decision policy for the storage device. Using this model, case studies are conducted for the two different systems. System load of these systems are peaked in the summer and winter, so these are analyzed and compared. In the third part, energy storage capacity value and expected profits are estimated when it provides energy, capacity, and frequency regulation services. To estimate capacity value, three steps approach is adopted. First, discretized stochastic dynamic programming is used to obtain decisions policies for the discretized states. These decision policies are used to get actual decisions by solving mixed-integer optimization in a rolling-horizon fashion. Then, capacity value of energy storage is estimated using simulation. A case (open full item for complete abstract)

Doctor of Philosophy, The Ohio State University, 2024, Industrial and Systems Engineering

Opaque models do not support Joint Activity and create brittle systems that fail rapidly when the model reaches the edges of its operating conditions. Instead, we should use models which are observable, directable, and predictable – qualities which are better suited by transparent or ‘explainable' models. However, using explainable models has traditionally been seen as a trade-off in machine performance, ignoring the potential benefits to the performance of the human machine teams. While the cost to model performance is negligible when considering the cost to the human machine team, there is a benefit to machine learning that has increased accuracy or capabilities when designed appropriately to deal with failure. Increased accuracy can indicate better alignment with the world and the increased capability to generalize across a broader variety of cases. Increased capability does not always have to come at the cost of explainability, and this dissertation will discuss approaches to make traditionally opaque models more usable in human machine teaming architectures.

Doctor of Philosophy, The Ohio State University, 2024, Industrial and Systems Engineering

This dissertation investigates the optimization of real-world energy applications, focusing on power and gas networks. These systems pose challenges due to their complex governing laws and practical considerations, leading to nonlinear optimization models that are often non-convex. Recent advancements in operations research, including relaxation and reformulation techniques, have provided avenues to tackle non-convexities. This dissertation delves into exponential conic programming (ECP) techniques, addressing two classical non-convex problems in power and gas networks. In power systems, the dissertation proposes a novel convexification technique through ECP reformulation to enhance system reliability. Similarly, in gas networks, this dissertation provides relaxation models based on ECP principles to solve signomial geometric programming (SGP) problems, demonstrating practicality and effectiveness. Through these efforts, the dissertation aims to advance optimization techniques for improving the reliability and operational efficiency of energy networks.

Master of Science (M.S.), University of Dayton, 2024, Chemical Engineering

There is a continuous thrust for cleaner and more sustainable alternatives for energy conversion with the increasing global energy demand. Among them, photovoltaics, specifically thin film solar cells are highly promising and are one of the fastest growing clean energy technologies in the United States. This research presents the synthesis and characterization of a set of novel multinary copper chalcogenide semiconductor nanocrystals (NCs), CuZn2ASxSe4-x consisting primarily of earth-abundant elements for applications in photovoltaic devices. A modified hot-injection method was used to synthesize these semiconductor NCs containing both S and Se chalcogens. The novelty of the new semiconductor NCs lies in the incorporation of multiple cations as well as two different chalcogen anions within the crystal lattice, which is an achievement from the materials synthesis aspect. The composition-controlled optical and photoluminescence properties of the CuZn2ASxSe4-x NCs were investigated via multi-modal material characterization including x-ray diffraction (XRD), ultraviolet-visible (UV-vis) spectroscopy, and photoluminescence spectroscopy (PL). The crystal structure, as determined from the XRD primarily consisted of the metastable wurtzite (P63mc) phase. The NCs exhibited direct band gap in the visible range that could be tuned both by varying the group III cation within the composition as well as the ratio of S/Se, based on the Tauc plot obtained from the UV-vis characterization. This work lays the groundwork for future investigations into the practical applications of copper chalcogenide NCs in solar energy conversion.

Doctor of Philosophy (Ph.D.), University of Dayton, 2024, Engineering

We describe a study on the effect of temperature variations on multi-channel Time to Digital Converters (TDC). The objective is to study the impact of ambient thermal variations on the performance of Field Programmable Gate Array (FPGA)-based Tapped Delay Line (TDL) TDC systems, while simultaneously meeting the requirements of high-precision time measurement, low-cost implementation, small size, and low power consumption. For our study we choose two devices, Xilinx Artix-7 and Microsemi ProASIC3L. The radiation-tolerant ProASIC3L device offers better stability in terms of thermal sensitivity and power consumption compared to the Artix-7. To assess the performance of the TDCs under varying thermal conditions, a laboratory thermal chamber was utilized to maintain ambient temperatures ranging from -75 to 80 °C. This analysis ensured a comprehensive evaluation of the TDCs performance across a wide operational range. By utilizing the Artix-7 and ProASIC3L devices, we achieved Root Mean Square (RMS) resolution of 24.7 and 554.59 picoseconds, respectively. We worked to determine the temperature sensitivity for both FPGA devices by observing a significantly low temperature coefficient using Artix-7, while temperature insensitive and stable performance are achieved using the ProASIC3L device. Total on-chip 3 power of 0.968 W was achieved using Artix-7 while less than 1.988 mW of power consumption was achieved using ProASIC3L device. The results and analysis presented in this study convince that the proposed design using the new generations of the FPGAs would help in the design and optimization of FPGA-based TDCs for many applications.

Master of Science in Engineering, Youngstown State University, 2024, Department of Mechanical, Industrial and Manufacturing Engineering

Lithium-ion batteries (LIBs) have been a vital technology since they were introduced to the world in the 1990s. Despite significant advancements in cost-effectiveness and production efficiency, there are still some obstacles that need to be addressed. Significantly, as lithium-ion battery (LIB) technology is increasingly used in the transportation industry to enable electric vehicles, the issue of industrial ethics and environmental sustainability becomes of extreme importance. We are currently developing water-based manufacturing procedures to achieve more environmentally friendly production of lithium-ion batteries. Our research focuses on analyzing the design elements and process dynamics involved in removing solvents from the electrode coatings of these batteries. We demonstrate the impact of substituting N-Methyl-2-pyrrolidone (NMP) with an aqueous solvent, specifically water, in the electrode. To describe the process of cathode drying, we employ a mathematical model at the continuous level. This model accounts for the simultaneous transmission of heat and mass, as well as phase change. The utilization of aqueous processing for electrode material has the potential of cost reduction and environmental effect reduction in existing lithium-ion battery (LIB) manufacturing processes. By substituting costly and hazardous binder solvents like N-methyl-2-pyrrolidone (NMP) with water-based processing, both material expenses and processing and capital equipment costs can be minimized. The optimization model will determine the most efficient factors for the energy consumption of the solvent drying process, which constitutes a significant component of the overall energy consumption in the drying process.

Doctor of Philosophy, The Ohio State University, 2024, Industrial and Systems Engineering

Decision-making problems in real life often involve uncertainties. One way to address such problems is to use stochastic optimization, where quantifying a probability distribution to represent the underlying uncertainty is critical. However, most often, only partial information about the uncertainty is available through a series of historical data and expert knowledge. This limitation becomes particularly significant if the decision maker is risk averse and needs to consider rare but high-impact events, for which the probability distribution cannot be accurately determined even with the available historical data. Distributionally Robust Stochastic Optimization (DRO) is an alternative approach that assumes that the underlying distribution is unknown but instead lies in an ambiguity set of distributions that is consistent with the available data. DRO then tries to optimize the worst-case expectation among all distributions in the ambiguity set. This dissertation focuses on effective scenarios in DROs defined using a finite number of realizations (also called scenarios) of the uncertain parameters. Effective scenarios are the critical scenarios in DRO in the sense that their removal alters the optimal objective function value. Ineffective scenarios, on the other hand, can be removed safely without changing the optimal value. In this dissertation, we investigate both the theoretical and computational aspects of effective scenarios. The first contribution of this dissertation links the effectiveness of a scenario to its worst-case distribution being always positive or uniquely zero under a general ambiguity set with finite support. We then narrow down our focus to DROs with ambiguity sets formed via the Cressie-Read power divergence family (DRO-CR) and the Wasserstein distance (DRO-W). This class of problems constitutes some of the most widely used DROs in the literature. We provide easy-to-check sufficient conditions to identify the effectiveness of scenarios fo (open full item for complete abstract)