https://etd.auburn.edu/handle/10415/2 2026-06-09T12:54:56Z https://etd.auburn.edu/handle/10415/10419 Reliability Analysis and Mechanical Characterization of Lead-Free Solders Using Interpretable Machine Learning Qasaimeh, Qais Predicting solder joint reliability and efficiently characterizing the mechanical properties of lead-free solder materials remain critical challenges in the electronics industry. Accurate reliability prediction is complicated by the large number of design and process parameters, as well as the complex nonlinear interactions among thermal, mechanical, and material-related factors that govern solder behavior under harsh operating conditions. Traditional reliability assessment methods, including analytical and numerical approaches, often rely on simplifying assumptions that limit their ability to capture these nonlinear relationships, thereby reducing prediction accuracy. At the same time, conventional mechanical characterization techniques, such as tensile, creep, and shear testing, are experimentally intensive, time-consuming, and resource-intensive. These challenges highlight the need for efficient, interpretable, and data-driven methodologies to improve both reliability prediction and mechanical characterization of lead-free solder materials. Recently, machine learning (ML) techniques have emerged as promising tools for addressing complex reliability and material characterization problems in electronics manufacturing. Their ability to model nonlinear relationships and incorporate numerous interacting variables has enabled more accurate prediction of solder joint behavior compared to conventional approaches. However, despite their predictive capabilities, many ML models suffer from limited interpretability, which reduces confidence in their predictions and limits their adoption in reliability-critical applications. Therefore, developing interpretable and robust data-driven frameworks is essential for enabling trustworthy decision-making in solder reliability assessment and material development To address these challenges, this dissertation develops a sequential data-driven framework through three interconnected studies. The first study establishes an interpretable reliability-prediction framework for solder joints subjected to thermal cycling by integrating linear mixed-effects (LME) models and artificial neural networks (ANNs) with sensitivity analysis. By leveraging both experimental and simulated reliability data, this study demonstrates how interpretable ML models can accurately predict solder joint life while simultaneously identifying the critical factors governing reliability behavior. The extracted insights align with empirical knowledge and experimental observations, providing a transparent foundation for data-driven reliability assessment. Building on this foundation, the second study addresses one of the major limitations in accelerated life testing (ALT): the handling of censored reliability data. In this study, advanced survival ML methods, including random survival forests (RSF), survival support vector machines (SSVM), and extreme gradient boosting (XGB), are employed and compared against traditional survival analysis models such as Weibull regression and the Cox proportional hazards model (Cox-PHM). Furthermore, Shapley additive explanations (SHAP)-based interpretability analysis is incorporated to explain the influence of input variables on reliability predictions. The results demonstrate that survival ML methods provide superior predictive performance and improved modeling of complex nonlinear reliability behavior under censored conditions. Finally, the third study extends these data-driven methodologies to efficiently characterize lead-free solder alloys by introducing a transferability-guided transfer learning (TL) framework for predicting tensile strength. In this framework, knowledge learned from data-rich conventional solder alloys is transferred to newly developed, data-scarce alloys, enabling accurate mechanical property prediction with limited experimental data. The proposed TL framework is benchmarked against multiple conventional ML approaches and demonstrates improved robustness and predictive accuracy in low-data scenarios. In addition, transferability and interpretability analyses provide insights into the effectiveness of knowledge transfer and the physical consistency of the learned relationships. Collectively, these three studies establish a comprehensive, interpretable, data-driven framework that progresses from solder joint reliability prediction to censored reliability modeling, and finally to mechanical property characterization and alloy development. The proposed methodologies accelerate the evaluation, optimization, and adoption of high-reliability lead-free solder materials in advanced electronic packaging applications. 2026-06-08T00:00:00Z https://etd.auburn.edu/handle/10415/10418 A Mixed-methods Multi-study Approach to Unveil the Implications of AI Integration for Employee Outcomes in the US Hotel Industry Bakir, Selim Artificial intelligence (AI) has emerged as a pivotal technology reshaping workplace dynamics and influencing employee behavior in the hotel industry. While existing research has largely emphasized operational efficiency and technological outcomes, a broader understanding of how employees perceive, experience, and respond to AI adoption remains limited. Addressing this gap, this dissertation employs a mixed-methods approach across three interrelated studies to examine the implications of AI for employee attitudes, behaviors, and managerial practices in the U.S. hotel industry. Drawing on the Transactional Theory of Stress and Coping, the Job Demands–Resources (JD–R) model, and Conservation of Resources (COR) theory, this research integrates quantitative and qualitative insights to uncover the psychological and organizational mechanisms through which AI shapes employee outcomes. Survey data collected from 413 hotel employees in the US reveal that different approaches to AI implementation produce divergent employee responses. Specifically, AI adoption is associated with lower levels of quiet quitting, whereas AI integration is linked to higher levels of withdrawal behaviors. Competitive psychological climate emerges as a key mediating mechanism, whereas technostress does not explain these relationships in this context. Complementing these findings, qualitative evidence from semi-structured interviews with 19 hotel managers demonstrates that AI adoption significantly influences employees’ career development, well-being, and work outcomes. Managers report shifts in employees’ career perceptions, including turnover intentions, job search behaviors, and career satisfaction, as well as both positive outcomes (e.g., improved efficiency) and negative outcomes (e.g., strain and techno-anxiety). Managerial support and transparent communication are identified as critical factors in mitigating adverse effects. Further quantitative analyses show that perceived AI job-replacement threat is positively associated with technostress and turnover intention, with technostress serving as a key mechanism linking AI-related threats to turnover intention and service sabotage. Collectively, this dissertation advances theoretical understanding of AI-enabled workplaces by clarifying how AI-related demands and resource threats influence employee withdrawal and deviant behaviors. The findings offer actionable insights for hospitality organizations seeking to implement AI in ways that support employee well-being and promote sustainable service performance. Keywords: AI Threat, Hotel Employees, Competitive Psychological Climate, Technostress, Quiet Quitting, Service Sabotage 2026-06-05T00:00:00Z https://etd.auburn.edu/handle/10415/10417 On Chemotaxis Model with Linear and Porous Medium Diffusion, Logistic Source and Consumption on \(\R^N\) Hassan, Zulaihat This dissertation is devoted to the study of chemotaxis systems with both linear diffusion and porous-medium-type diffusion, logistic source terms, and consumption of a chemical substance on $\mathbb{R}^N$. Chemotaxis systems are mathematical models describing the aggregation of cells driven by their directed movement in response to gradients of chemicals in their environment, which may act as attractants or repellents. In the first part of this dissertation, we investigate a chemotaxis model with linear diffusion. We study fundamental problems such as the local and global existence of classical solutions with nonnegative initial data, which may be integrable or non-integrable. Under suitable smallness assumptions on the product of the initial chemical concentration and the chemotactic sensitivity, we prove the existence of a unique global classical solution. For non-integrable initial data, we develop a novel weighted energy method to establish global existence and boundedness. By introducing carefully chosen cut-off functions, we localize $L^p$-estimates uniformly in space. This approach extends known results for bounded domains and is applicable to other chemotaxis systems. We also study the stability of strictly positive solutions and the spreading behavior of solutions with compactly supported initial data. We show that the chemical does not, in general, hinder the spreading of the species, and it does not accelerate the spreading speed when the initial chemical concentration decays spatially or in the chemorepellent case with small sensitivity. Numerical simulations further reveal a phase transition in the sensitivity $\chi$: when the chemical is initially uniformly distributed in space, acceleration occurs only when $\chi$ exceeds a critical positive value. In the second part, we study the local and global existence of weak solutions for the porous-medium diffusion case. For general bounded, possibly non-integrable initial data, we prove the existence of global weak solutions that remain uniformly bounded for all time. The proof is based on local $L^p$ estimates, uniform in time, obtained through a new continuity-type argument combined with Moser iteration to derive $L^\infty$ bounds. We also investigate regularity and prove uniqueness of weak solutions for sufficiently smooth initial data under suitable conditions on the diffusion exponent. 2026-06-01T00:00:00Z https://etd.auburn.edu/handle/10415/10416 Tube MPC for Robust Lateral Control of a Class 8 Tractor-Trailer with Parameter Uncertainty Ellison, Evan This thesis presents a Model Predictive Control (MPC) design for motion planning and control of a five axle tractor-trailer vehicle. The targeted use case for this design is for Society of Automotive Engineers (SAE) Level 3-4 features which may include automated highway driving and lane change or obstacle avoidance maneuvers. Autonomous control of commercial tractor trailer vehicles, specifically class 8 trucks, presents unique challenges due to the need for specific safety guarantees and lack of accurate knowledge of all of the model parameters, such as the mass and yaw inertia of the payload in the trailer. These challenges can be handled in part by MPC due to its ability to enforce constraints and find optimal trajectories with respect to an objective. Additionally, many techniques for ensuring constraints are satisfied under uncertainty exist which can prove useful for this application. In this thesis, a commonly used dynamic model for tractortrailers is first presented. Next, a full prediction model for use in the MPC is developed, which combines the equations for propagating position with respect to the road and a model for the steering actuator with the lateral dynamic model of the vehicle. An MPC design is then introduced by defining the optimal control problem and solving it as a Quadratic Program (QP). The MPC is able to plan and execute a trajectory that ensures constraints related to the vehicle’s position and trailer states such as the hitch angle can be met. A higher update rate feedback controller is used to aid the tracking of the latest solution between MPC updates. A constraint tightening technique is also applied which constructs an error tube around the planned trajectory based on the uncertainty of model parameters. An analysis of the total accuracy and performance in different scenarios is presented. The combined online planning and control scheme is validated in simulation, and the MPC performance with and without constraint tightening is compared for several relevant scenarios, and improvements in the number of scenarios that satisfy the lateral position and hitch angle constraints is demonstrated. Finally, the real-world capability of the design is demonstrated on an autonomous Peterbilt 579 with a trailer attached. The experimental testing demonstrates the feasibility of the concept for real-time control through lane keeping tests, resulting in absolute tracking errors with at most a mean of 18.9 cm and a standard deviation of 11.9 cm. 2026-05-19T00:00:00Z