Permanent deformation in the form of creep is added to a one-dimensional model of a radio-frequency micro-electro-mechanical system (RF-MEMS). Due to uncertainty in the material property values, calibration under uncertainty is carried out through comparison to experiments in order to determine appropriate boundary conditions and material property values. Further uncertainty in the input parameters, in the form of probability distribution functions of geometric device properties, is included in simulations and propagated to the device performance as a function of time. The effect of realistic power-law grain size distributions on the creep response of thin RF-MEMS films is examined through the use of a finite volume software suite designed for the computational modelling of MEMS. It is seen that the use of a realistic height-dependent power-law distribution of grain sizes in the film in place of a uniform grain size has the effect of increasing the simulated creep rate and the uncertainty in its value. The effect is seen to be the result of the difference between the model with a homogeneous grain size and the model with a non-homogeneous grain size. Realistic variations in the grain size distribution for a given film are seen to have a smaller effect. Finally, in order to incorporate variations in thickness in manufactured devices, variation in the thickness of the membrane across the length and width is considered in a 3D finite element model, and variation of thickness along the length is added to the earlier one-dimensional RF-MEMS model. Estimated uncertainty in the film profile is propagated to selected device performance metrics. The effect of film thickness variation along the length of the film is seen to be greater than the effect of variation across the width.

All quantitative measurements have a degree of measurement uncertainty. While the term uncertainty can be essentially defined as doubt, measurement uncertainty in this sense instead inspires assurance in a quantitative value to a certain degree of confidence. Dating back to the advent of modern statistics in the 1700s, an international consensus on measurement uncertainty did not occur until the 1995 release of the Guide to the Expression of Uncertainty in Measurement (GUM), the fundamental document on the subject. The GUM was further adopted by major players in the field of measurement including the International Bureau of Weights and Measures (BIPM), National Measurement Institutes (NMI), and the International Organization for Standardization (ISO), and is used as the gold standard of documentary standards in labs around the country. Gunshot residue (GSR) patterns of distribution are used to establish a range of possible distances that the muzzle of the firearm was from the target in order to piece together a particular series of events. Using the firearm and ammunition that was involved in that particular crime, an analyst can perform test fires using fabric swatches attached to test panels at varying muzzle-to-target distances, generally every three to six inches between contact and 48 inches. This allows for the creation of comparable patterns of soot and GSR to the actual pieces of evidence. Because different distances can have considerably different residue patterns, it is important that a method for creating the test panels minimize uncertainty in order to be considered reliable and reproducible. When establishing a protocol for determining the measurement uncertainty in the creation of test panels, the two most important factors are the measuring device and a repeatability study. A measuring device, in this case a stainless-steel ruler, with metrological traceability reduces the measurement uncertainty because every value is reliable and traceable back to an original source. A repeatability study is then used to take numerous measurements over time under similar conditions. Using this data, statistical analysis can be applied to evaluate the standard deviations and uncertainties. A total of 238 measurements were taken by eight members of the Boston Police Department Crime Laboratory on eleven different days over the course of a month. The measurements were divided into eight baseline distances that the firing device, a Ransom Rest, had been set to: 3”, 6”, 9”, 12”, 18”, 24”, 36”, and 42”. The data was analyzed as a whole, as well as split into two groups: a group of four analysts who are proficient and authorized to perform GSR distance determination testing (Group A), and a second group of four analysts with no GSR distance determination training or experience (Group B). At a confidence interval of 95.45%, the reported uncertainty was found to be 0.082 inches for the total group, 0.045 inches for the group trained in performing GSR distance determination, and 0.043 inches for the group with no experience in distance determination testing. F-test statistical analysis of the standard deviations of each distance, along with a comparison of the uncertainties, indicates no significant difference between the abilities of the two groups and that it’s possible a new uncertainty of measurement will not be required when current GSR distance determination analysts leave or new analysts are hired and trained, given that all other variables remain constant. The outlined method and experiment for determining measurement uncertainty was successful in that it met the four main requirements set forward by the American National Standards Institute (ANSI) National Accreditation Board (ANAB): (a) include the specific measuring device or instrument used for a reported test result in the estimation of measurement uncertainty for that test method; (b) include the process of rounding the expanded uncertainty; (c) require the coverage probability of the expanded uncertainty to be a minimum of 95.45%; and (d) specify a schedule to review and/or recalculate the measurement uncertainty.

The increasing need for groundwater as a source for fresh water and the continuous deterioration in many places around the world of that precious source as a result of anthropogenic sources of pollution highlights the need for efficient groundwater resources management. To be efficient, groundwater resources management requires efficient access to reliable information that can be acquired through monitoring. Due to the limited resources to implement a monitoring program, a groundwater quality monitoring network design should identify what is an optimal network from the point of view of cost, the value of information collected, and the amount of uncertainty that will exist about the quality of groundwater. When considering the potential social impact of monitoring, the design of a network should involve all stakeholders including people who are consuming the groundwater. This research introduces a methodology for groundwater quality monitoring network design that utilizes state-of-the-art learning machines that have been developed from the general area of statistical learning theory. The methodology takes into account uncertainties in aquifer properties, pollution transport processes, and climate. To check the feasibility of the network design, the research introduces a methodology to estimate the value of information (VOI) provided by the network using a decision tree model. Finally, the research presents the results of a survey administered in the study area to determine whether the implementation of the monitoring network design could be supported. Applying these methodologies on the Eocene Aquifer, Palestine indicates that statistical learning machines can be most effectively used to design a groundwater quality monitoring network in real-life aquifers. On the other hand, VOI analysis indicates that for the value of monitoring to exceed the cost of monitoring, more work is needed to improve the accuracy of the network and to increase people’s awareness of the pollution problem and the available alternatives.

A vast literature has investigated the returns to investment in agriculture research and generally found extremely high rates of return. These results suggest policymakers would do well to maintain or increase resource allocation to public agricultural research. Remarkably little attention has been paid, however, to the issue of how best to allocate public agricultural research funding between competing research areas and organizations. This paper considers the relative returns to alternative uses of public agricultural research funds committed to the agricultural experiment stations of 10 western states of the United States over the years 1967-91. A model of expected utility maximization subject to risk is presented with comparative analysis. After establishing empirically that the mean variance analysis would be an inappropriate method to solving the problem, a stochastic dominance testing method is employed to identify dominated and undominated research categories and state agricultural experiment stations. The mean variance analysis also is used to evaluate whether research productivity has been increasing or decreasing over time, and to establish which among the western states hold absolute advantage in particular research areas.

The existing literature determining the optimal rotation period of a forest stand under conditions of certainty, as well as under uncertainty, lacks the genera 1 scope to be useful. A forest provides timber of commercial value, a flow of recreational services, and other valuable environmental s ervices. Providing goods and services invo lves benefits as well as costs. Relevant management decisions depend on the net va 1 ues that can be obtai ned. The present work developes a more general model for determining an optimal rotation period incorporating various fixed and variable costs associated with timber production and recreational services in an environment of certainty and uncertainty. It is shown that under certainty, the optimal rotation period is likely to be finite and depending on the values of benefits and costs the rotation period indicated by the solution of this model may be identical to, shorter, or longer than that indicated by a model ignoring net values. In addition, a generalized Faustmann rule under certainty (when only recreational value i s added to the model ) using optimal control (maximum principle) as the analytical tool has been developed and the impact of two sources of uncertainties on the optimal rotation decision in the context of the more generalized model is analyzed. They are (1) uncertainty related to future stumpage price, and (2) uncertainty related to the future stock of trees due to unpredictable natural catastrophes. Under price uncertainty the optimal rotation period will be longer than that under conditions of certainty if the forest operator is risk averse. In addition, the period wi ll be lengthened with increasing risk and shortened with increasing expected stumpage price under noni ncreasing absolute risk aversion of the forest operator. The risk of catastrophic destruction of the biomass whether total or partial will lead to a rotation period dependent on the value of the average rate of occurrence of catastrophes.

Environmental scientists increasingly use remotely-sensed images to measure how rivers develop over time and respond to upstream changes in environmental drivers such as land use, urbanization, deforestation and agricultural practices. These measurements are subject to uncertainty that can bias conclusions. The first step towards accurate interpretation of river channel change is properly quantifying and accounting for uncertainty involved in measuring changes in river morphology. In Chapter 2 we develop a comprehensive framework for quantifying uncertainty in measurements of river change derived from aerial images. The framework builds upon previous uncertainty research by describing best practices and context-specific strategies, comparing each approach and outlining how to best handle measurements that fall below the minimum level of detection. We use this framework in subsequent chapters to reduce the impact of erroneous measurements. Chapter 3 evaluates how the time interval between aerial images influences the rates at which river channels appear to laterally migrate across their floodplains. Multiple lines of evidence indicate that river migration measurements obtained over longer time intervals (20+ years) will underestimate the ‘true’ rate because the river channel is more likely to have reversed the direction of migration, which erases part of the record of gross erosion as seen from aerial images. If the images don’t capture channel reversals and periodic episodes of fast erosion, the river appears to have migrated a shorter distance (which corresponds to a slower rate) than reality. Obtaining multiple measurements over shorter time intervals (< 5 years) and limiting direct comparisons to similar time intervals can reduce bias when inferring how river migration rates may have changed over time. Chapter 4 explores the physical processes governing the relationship between river curvature and the rate of river migration along a series of meander bends. We used fine-scale empirical measurements and geospatial analyses to confirm theory and models indicating that migration and curvature exhibit a monotonic relationship. The results will improve models seeking to emulate river meander migration patterns.

This work presents the state of the art in hierarchically decomposed multilevel optimization. This work is expanded with the inclusion of evidence theory with the multilevel framework for the quantification of epistemic uncertainty. The novel method, Evidence-Based Multilevel Design optimization, is then used to solve two analytical optimization problems. This method is also used to explore the effect of the belief structure on the final solution. A methodology is presented to reduce the costs of evidence-based optimization through manipulation of the belief structure. In addition, a transport aircraft wing is also solved with multilevel optimization without uncertainty. This complex, real world optimization problem shows the capability of decomposed multilevel framework to reduce costs of solving computationally expensive problems with black box analyses.

The primary goal of this research is to investigate the mechanical reinforcing efficiencies of carbon nanofibers in a thermoset polymer material (vinyl ester), considering the presence of the three-dimensional interphase region between nanofiber and matrix, as well as the waviness of the nanofibers. The elasticity-driven response (buckling) and energy absorption efficiency (crush performance) of the structures made of those composites are investigated. The structural/material optimization problem is solved for both buckling and energy absorption. Due to the nondeterministic nature of the influential parameters (fiber, matrix, and interphase geometric and material properties) on the overall properties of the composite, this study considers the probabilistically distributed random variables associated with the material constituents. The uncertainties associated with the material constituents are propagated to the overall properties of the composite material as well as the performance of composite structures made of such nanocomposites. Finally, the design optimization of a composite structure under uncertainty of material constituents is performed for both buckling and energy absorption as structural performance.

As online shopping continues to grow rapidly, research indicates its massive uptake can be the result of the integration of social media technologies within the e-commerce interface. This has been further exacerbated by the COVID-19 pandemic which has led to an acceleration in the use of e-commerce. However, despite the growing popularity of e-commerce, shopping online is characterized by high levels of uncertainty given the spatial and temporal separation between consumer and vendor. This presents a dire impact on a consumer’s decision-making process and can specifically impact a consumer’s decision confidence. Decision confidence is an important construct that has been shown to be central in influencing a consumer’s behaviour, specifically in regards to purchase-related activities. Few studies have shed light on the factors influencing a consumer’s decision confidence in the e-commerce context. Online shopping platforms fitted with social data markers are able to gauge and track the activities and attributes of online consumers, providing convenient heuristics on various measures such as the total number of recommendations for a product, or the degree of similarity between consumers. These markers may facilitate group identification through the development of herd behaviour. Herd behaviour arises in situations of uncertainty and motivates individuals to identify with a group (herd), and conform to its actions. Various aspects of a herd can influence group identification. This research focuses on two aspects of herd behaviour in e-commerce environments: homophily and herd size. Homophily is the degree to which individuals are similar, and in this study, we examine homophily from the perspective of an individual and the herd to which they may be exposed to. Herd size is the number of individuals in a group taking a specific action such as an online purchase decision. Drawing on the social identity approach and uncertainty identity theory, this investigation hones in on how homophily and herd size arise in the e-commerce context, and examines how group identification through homophily and herd size may reduce uncertainty and build decision confidence through the formation of trust, entitativity, sense of community and information helpfulness. A research model is developed along with a set of supported hypotheses. An online experiment utilizing a hypothetical e-commerce website was conducted with 400 participants. The results were analyzed using structural equation modeling and choice-based conjoint analysis. The results suggest that while homophily significantly impacts trust, sense of community, entitativity and information helpfulness, herd size does not. The results also suggest that whereas trust, sense of community and information helpfulness positively impact decision confidence, entitativity does not. It was also empirically demonstrated that participants preferred measures of homophily in the e-commerce interface more than measures of herd size. Furthermore, it was demonstrated that participants preferred measures indicating homophily in interests and demographics more than information on either alone, and that information on homophily in interests was more preferred than information on homophily in demographics. / Dissertation / Doctor of Philosophy (PhD) / Despite the rapid uptake in e-commerce, shopping online continues to be characterized by high levels of uncertainty given the spatial and temporal separation between consumer and vendor. This uncertainty negatively impacts a consumer’s decision confidence, which is a key driver in influencing consumer behaviour. Drawing on the social identity approach and uncertainty identity theory, this study investigates how elements of the e-commerce interface can facilitate group identification, a process which has been empirically demonstrated to reduce uncertainty and thereby increase decision confidence. Findings of this study suggest that measures of homophily embedded within an e-commerce platform can work through various mediators to facilitate group membership which can positively impact a consumer’s decision confidence. Theoretical and practical contributions of this study are discussed for researchers, academics, and practitioners wishing to explore those aspects of online shopping that an aid the decision-making process through group-related processes.

This text is a written articulation of my MFA thesis show entitled Corporeal Thresholds. It aims to share the poignant moments that inspired the work and contextualizes my practice within the framework of metaphysics and the phenomenology of perception. Specific topics include the body and the visceral, doubt and certainty, love and loss, and the defining spaces that influence our lives.