No description available.

In this thesis, we investigate the problem of fault tolerance in the framework of discrete-event systems (DES). We introduce our setting, and then provide a set of fault-tolerant definitions designed to capture different types of fault scenarios and to ensure that our system remains controllable and nonblocking in each scenario. This is a passive approach that relies upon inherent redundancy in the system being controlled, and focuses on the intermittent occurrence of faults. Our approach provides an easy method for users to add fault events to a system model and is based on user designed supervisors and verification. As synthesis algorithms have higher complexity than verification algorithms, our approach should be applicable to larger systems than existing active fault-recovery methods that are synthesis based. Also, modular supervisors are typically easier to understand and implement than the results of synthesis. Finally, our approach does not require expensive (in terms of algorithm complexity) fault diagnosers to work. Diagnosers are, however, required by existing methods to know when to switch to a recovery supervisor. As a result, the response time of diagnosers is not an issue for us. Our supervisors are designed to handle the original and the faulted system. In this thesis, we next present algorithms to verify these properties followed by complexity analyses and correctness proofs of the algorithms. Finally, examples are provided to illustrate our approach. In the above framework, permanent faults can be modelled, but the current method was onerous. To address this, we then introduce a new modeling approach for permanent faults that is easy to use, as well as a set of new permanent fault-tolerant definitions. These definitions are designed to capture several types of permanent fault scenarios and to ensure that our system remains controllable and nonblocking in each scenario. New definitions and scenarios were required as the previous ones were incompatible with the new permanent fault modeling approach. We then present algorithms to verify these properties followed by complexity analyses and correctness proofs of the algorithms. An example is then provided to illustrate our approach. Finally, we extend the above intermittent and permanent fault-tolerant approach to the timed DES setting. As before, we introduced new fault-tolerant properties and algorithms. We then provide complexity analyses and correctness proofs for the algorithms. An example is then provided to illustrate our approach. / Thesis / Doctor of Philosophy (PhD)

Yes / Researchers have used the simulation technique to develop new networks and test, modify, and optimize existing ones. The scientific community has developed a wide range of network simulators to fulfil these objectives and facilitate this creative process. However, selecting a suitable simulator appropriate for a given purpose requires a comprehensive study of network simulators. The current literature on network simulators has limitations. Limited simulators have been included in the studies with functional and performance criteria appropriate for comparison not been considered, and a reasonable selection model for selecting the suitable simulator has not been presented. To overcome these limitations, we studied twenty-three existing network simulators with classifications, additional comparison parameters, system limitations, and comparisons using several criteria. / This work was supported by the Petroleum Technology Development Fund (PTDF) Nigeria with grant number PTDF/ED/PHD/MAS/179/17.

This research presents results from a multi-site field study of 51 Kaizen event teams in six manufacturing organizations. Although Kaizen events have been growing in popularity since the mid 1990s, to date, there has been no systematic empirical research on the determinants of Kaizen event effectiveness. To address this need, a theory-driven model of event effectiveness is developed, drawn from extant Kaizen event practitioner articles and related literature on projects and teams. This model relates Kaizen event outcomes to hypothesized key input factors and hypothesized key process factors. In addition, process factors are hypothesized to partially mediate the relationship between input factors and outcomes. Following sociotechnical systems (STS) theory, both technical and social (human resource) aspects of Kaizen event performance are measured. Relationships between outcomes, process factors and input factors are analyzed through regression, using generalized estimating equations (GEE) to account for potential correlation in residuals within organizations. The research found a significant positive correlation between the two social system outcomes (attitude toward Kaizen events and employee gains in problem-solving knowledge, skills and attitudes). In addition, the research found significant positive correlations between the social system outcomes and one technical system outcome (team member perceptions of the impact of the Kaizen event on the target work area). However, none of the three technical system outcomes (employee perceptions of event impact, facilitator ratings of event success and actual percentage of team goals achieved) were significantly correlated. In addition, the research found that each outcome variable had a unique set of input and process predictors. However, management support and goal difficulty were a common predictors of three out of five outcomes. Unexpected findings include negative relationships between functional diversity, team and team leader Kaizen event experience, and action orientation and one or more outcomes. However, many of the findings confirmed recommendations in Kaizen event practitioner articles and the project and team literature. Furthermore, support for the mediation hypothesis was found for most outcome measures. These findings will be useful both for informing Kaizen event design in practicing organizations and for informing future Kaizen event research. / Ph. D.

Throughout the past several decades, research regarding counselor resilience has shifted from a pathology-based to a strengths-based approach. As a result, researchers have moved away from primarily identifying risk factors and now focus on protective factors. Researchers have found that social supports serve as a protective factor in counselor resilience. However, there is a lack of understanding of how counselors receive that social support, specifically after a professional adverse event. Professional adverse events are common given the nature of counseling work. For example, undesirable occurrences such as client suicide, attempted suicide, life threatening illnesses, accidents, overdose, or loss of a child are considered professional adverse events. This study explored how ten professional counselors experienced social support following professional adverse events. Three themes emerged from these counselors' stories: difficulty seeking support, misplaced support, and acts of kindness. Implications for counselors include honoring both confidentiality and their own humanness, the cultivating co-regulating relationships, and reinforcing acts of kindness. Counselor educators and supervisors can foster counselor resilience by using the implications to teach counselors how to invite effective social support. / Doctor of Philosophy / Being a counselor can be both challenging and rewarding. Given the heavy caseloads and complexity of cases, it is common for counselors to experience adverse professional events. Research shows that protective factors serve as a buffer against stress. Social support is a protective factor that assists counselors in maintaining wellness and building resiliency. This study explored how professional counselors received support from interpersonal relationships following a professional adverse event. Findings from this study indicate the importance of counselors honoring their own humanness while protecting the client’s confidentiality, the importance of having co-regulating relationships, and the healing power of acts of kindness.

Neural Networks have proven themselves very adept at solving a wide variety of problems, in particular they accel at image processing. However, it remains unknown how well they perform under memory errors. This thesis focuses on the robustness of neural networks under memory errors, specifically single event upset style errors where single bits flip in a network's trained parameters. The main goal of these experiments is to determine if different neural network architectures are more robust than others. Initial experiments show that MLPs are more robust than CNNs. Within MLPs, deeper MLPs are more robust and for CNNs larger kernels are more robust. Additionally, the CNNs displayed bimodal failure behavior, where memory errors would either not affect the performance of the network, or they would degrade its performance to be on par with random guessing. VGG16, ResNet50, and InceptionV3 were also tested for their robustness. ResNet50 and InceptionV3 were both more robust than VGG16. This could be due to their use of Batch Normalization or the fact that ResNet50 and InceptionV3 both use shortcut connections in their hidden layers. After determining which networks were most robust, some estimated error rates from neutrons were calculated for space environments to determine if these architectures were robust enough to survive. It was determined that large MLPs, ResNet50, and InceptionV3 could survive in Low Earth Orbit on commercial memory technology and only use software error correction. / Master of Science / Neural networks are a new kind of algorithm that are revolutionizing the field of computer vision. Neural networks can be used to detect and classify objects in pictures or videos with accuracy on par with human performance. Neural networks achieve such good performance after a long training process during which many parameters are adjusted until the network can correctly identify objects such as cats, dogs, trucks, and more. These trained parameters are then stored in a computers memory and then recalled whenever the neural network is used for a computer vision task. Some computer vision tasks are safety critical, such as a self-driving car’s pedestrian detector. An error in that detector could lead to loss of life, so neural networks must be robust against a wide variety of errors. This thesis will focus on a specific kind of error: bit flips in the parameters of a neural networks stored in a computer’s memory. The main goal of these bit flip experiments is to determine if certain kinds of neural networks are more robust than others. Initial experiments show that MLP (Multilayer Perceptions) style networks are more robust than CNNs (Convolutional Neural Network). For MLP style networks, making the network deeper with more layers increases the accuracy and the robustness of the network. However, for the CNNs increasing the depth only increased the accuracy, not the robustness. The robustness of the CNNs displayed an interesting trend of bimodal failure behavior, where memory errors would either not affect the performance of the network, or they would degrade its performance to be on par with random guessing. A second set of experiments were run to focus more on CNN robustness because CNNs are much more capable than MLPs. The second set of experiments focused on the robustness of VGG16, ResNet50, and InceptionV3. These CNNs are all very large and have very good performance on real world datasets such as ImageNet. Bit flip experiments showed that ResNet50 and InceptionV3 were both more robust than VGG16. This could be due to their use of Batch Normalization or the fact that ResNet50 and InceptionV3 both use shortcut connections within their network architecture. However, all three networks still displayed the bimodal failure mode seen previously. After determining which networks were most robust, some estimated error rates were calculated for a real world environment. The chosen environment was the space environment because it naturally causes a high amount of bit flips in memory, so if NASA were to use neural networks on any rovers they would need to make sure the neural networks are robust enough to survive. It was determined that large MLPs, ResNet50, and InceptionV3 could survive in Low Earth Orbit on commercial memory technology and only use software error correction. Using only software error correction will allow satellite makers to build more advanced satellites without paying extra money for radiation-hardened electronics.

This paper applies an event study approach to measure the impact of United States Department of Agriculture (USDA) reports on dairy futures price volatility over January 2011 to December 2023. Dairy futures are a relatively understudied commodity market with a unique pricing structure and settlement procedure. An E-GARCH model is used to estimate price volatility with exogenous dummy variables of lagged volume, NDPSR, WASDE, Cold Storage, Dairy Products, and Milk Production. Milk Production had the strongest impact, significantly increasing price volatility in all markets but Class III. National Dairy Product Sales Report (NDPSR) was found to significantly decrease volatility in all markets except Class III. The other reports studied had mixed impacts on the dairy markets. / Master of Science / This paper attempts to measure the impact of United States Department of Agriculture (USDA) reports on dairy futures price volatility from January 2011 to December 2023. Dairy futures are a relatively understudied commodity market with a unique pricing structure and settlement procedure. A volatility estimation model is used to estimate price volatility with exogenous variables of lagged volume, NDPSR, WASDE, Cold Storage, Dairy Products, and Milk Production. Milk Production is the most impactful of USDA reports, positively impacting all markets but Class III. National Dairy Product Sales Report (NDPSR) was found to have a negative impact on all markets except Class III. The other reports studied had mixed impacts on markets.

This dissertation consists of two essays that focus on the topics related to asset pricing with jump risks. The first essay explores the effect of disaster risk on the beliefs and portfolio choices of ambiguity-averse agents. With the introduction of Cressie-Read discrepancies, a time-varying pessimism state variable arises endogenously, generating time-varying disaster risk. In the event of a disaster, agents heighten their pessimism, anticipating subsequent disasters to arrive sooner. Within this framework, we deduce optimal consumption and portfolio choices that are robust to model misspecification. Additionally, our measure of pessimism aids in understanding the stylized facts derived from Vanguard’s retail investor survey data, as reported in Giglio et al. (2021). In the second essay, I construct a novel measure to assess the impact of macro announcements on investors’ risk expectations using S&P 500 index and Treasury futures options. This measure corrects the systematic downward jumps in the option- implied variance measure and isolates innovations of investors’ risk expectations after macro-announcements. Applied to key economic releases, including FOMC meetings, GDP, PPI, and Employment data announcements, this measure reveals that macro announcements significantly increase investors’ risk expectations compared to pre-announcement levels. Furthermore, I show investor sentiment significantly declines following macro-announcements with heightened risk expectations, and tail risk positively correlates with risk expectations.

The rise of big data, especially social media data (e.g., Twitter, Facebook, Youtube), gives new opportunities to the understanding of human behavior. Consequently, novel computing methods for mining patterns in social media data are therefore desired. Through applying these approaches, it has become possible to aggregate public available data to capture triggers underlying events, detect on-going trends, and forecast future happenings. This thesis focuses on developing methods for social media analysis. Specifically, five directions are proposed here: 1) semi-supervised detection for targeted-domain events, 2) topical interaction study among multiple datasets, 3) discriminative learning about the identifications for common and distinctive topics, 4) epidemics modeling for flu forecasting with simulation via signals from social media data, 5) storyline generation for massive unorganized documents. / Ph. D. / The rise of “big data”, especially social media data (e.g., Twitter, Facebook, Youtube), gives new opportunities to the understanding of human behavior. Consequently, novel computing methods for mining patterns in social media data are therefore desired. Through applying these approaches, it has become possible to aggregate public available data to capture triggers underlying events, detect on-going trends, and forecast future happenings. This dissertation provides comprehensive studies for social media data analysis. The goals of the dissertation include: event early detection, future event prediction, and event chain organization. Specifically, these goals are achieved through efforts in the following aspects: (1) semi-supervised and unsupervised methods are developed to collect early signals from social media data and detect on-going events; (2) graphical models are proposed to model the interaction and comparison among multiple datasets; (3) traditional computational methods are combined with new emerge social media data analysis for the purpose of fast epidemic prediction; (4) events in different time stamps are organized into event chains via novel probabilistic models. The effectiveness of our approaches is evaluated using various datasets, such as Twitter posts and news articles. Also, interesting case studies are provided to show models’ abilities in the real world exploration.

Thesis advisor: Jane Flanagan / Background: Preventable adverse events are an unfortunately frequent occurrence of pediatric health care. Disclosure of preventable adverse events to patients is a vital aspect of ethical and just practice. Pediatric nurses’ have a unique role as part of the clinical care team. Despite the prevalence of preventable adverse events and the impact of nurses, best practice for pediatric nurses during disclosure is not specified. In addition, it is unclear how pediatric patient and their family. This work provides a foundation for future nursing research and the development and identification of best practice for pediatric preventable adverse event disclosure. Methods: First, thorough review of existing literature identified gaps and key themes. Secondly, a cross-sectional survey shared via social media provided insight into the current policy, education, and pediatric nurses’ involvement in PAE. Lastly, pediatric nurses’ perspectives were unveiled via narrative interviews, adding the voice of nurses into the dialogue. Results: Pediatric nurses in the U.S. want the option to be present during disclosure to patients and their families. Currently, nurses are seldom present during disclosure and do not routinely receive disclosure training, nor do they have a policy to guide them through the process. While there has been a trend towards the use of interdisciplinary disclosure teams, it is unclear what role a nurse has. Conclusions: This exploratory work is foundational to understanding pediatric nurses and PAE disclosure and future research exploring best practice for policy, education, and practice are needed. / Thesis (PhD) — Boston College, 2024. / Submitted to: Boston College. Connell School of Nursing. / Discipline: Nursing.