Theses and Dissertations from UMD
Permanent URI for this communityhttp://hdl.handle.net/1903/2
New submissions to the thesis/dissertation collections are added automatically as they are received from the Graduate School. Currently, the Graduate School deposits all theses and dissertations from a given semester after the official graduation date. This means that there may be up to a 4 month delay in the appearance of a give thesis/dissertation in DRUM
More information is available at Theses and Dissertations at University of Maryland Libraries.
Browse
5 results
Search Results
Item A SYSTEMS RELIABILITY APPROACH TO MODELING OPERATIONAL RISKS IN COMPLEX ENGINEERED SYSTEMS(2018) Komey, Adiel Ayi; Baecher, Gregory B; Modarres, Mohammad; Civil Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Since the beginning of the industrial revolution in the late 18th century, the cause of many serious accidents in hydrosystems engineering has shifted from natural causes to human and technology related causes as these systems get more complex. While natural disasters still account for a significant amount of human and material losses, man-made disasters are responsible for an increasingly large portion of the toll, especially in the safety critical domain such as Dam and Levee systems. The reliable performance of hydraulic flow-control systems such as dams, reservoirs, levees etc. depends on the time-varying demands placed upon it by hydrology, operating rules, the interactions among subsystem components, the vagaries of operator interventions and natural disturbances. In the past, engineers have concerned themselves with understanding how the component parts of dam systems operate individually and not how the components interact with one another. Contemporary engineering practices do not address many common causes of accidents and failures, which are unforeseen combinations of usual conditions. In recent decades, the most likely causes of failures associated with dams have more often had to do with sensor and control systems, human agency, and inadequate maintenance than with extreme loads such as floods and earthquakes. This thesis presents a new approach, which combines simulation, engineering reliability modeling, and systems engineering. The new approach seeks to explore the possibilities inherent in taking a systems perspective to modeling the reliability of flow-control functions in hydrosystems engineering. Thus, taking into account the interconnections and dependencies between different components of the system, changes over time in their state as well as the influence upon the system of organizational limitations, human errors and external disturbances. The proposed framework attempts to consider all the physical and functional interrelationships between the parts of the dam and reservoir, and to combine the analysis of the parts in their functional and spatial interrelationships in a unified structure. The method attempts to bring together the systems aspects of engineering and operational concerns in a way that emphasizes their interactions. The argument made in this thesis is that systems reliability approach to analyzing operational risks—precisely because it treats systems interactions—cannot be based on the decomposition, linear methods of contemporary practice. These methods cannot logically capture the interactions and feedback of complex systems. The proposed systems approach relies on understanding and accurately characterizing the complex interrelationships among different elements within an engineered system. The modeling framework allows for analysis of how structural changes in one part of a system might affect the behavior of the system as a whole, or how the system responds to emergent geophysical processes. The implementation of the proposed approach is presented in the context of two case studies of US and Canadian water projects: Wolf Creek Dam in Kentucky and the Lower Mattagami River Project in Northern Ontario.Item Statistical Inference Using Data From Multiple Files Combined Through Record Linkage(2018) HAN, YING; Lahiri, Partha; Mathematical Statistics; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Record linkage methods help us combine multiple data sets from different sources when a single data set with all necessary information is unavailable or when data collection on additional variables is time consuming and extremely costly. Linkage errors are inevitable in the linked data set because of the unavailability of an error-free and unique identifier and because of possible errors in measuring or recording. It has been realized that even a small amount of linkage errors can lead to substantial bias and increase variability in estimating the parameters of a statistical model. The importance of incorporating uncertainty of the record linkage process into the statistical analysis step cannot be overemphasized. The current research is mainly focused on the regression analysis of the linked data. The record linkage and statistical analysis processes are treated as two separate steps. Due to the limited information about the record linkage process, simplifying assumptions on the linkage mechanism have to be made. In reality, however, these assumptions may be violated. Also, most of the existing linkage error models are built on the linked data set, which only contains records for the designated links. Information about linkage errors carried by the designated non-links is missing. In the dissertation, we provide general methodologies for both regression analysis and small area estimation using data from multiple files. A general integrated model is proposed to combine the record linkage and statistical analysis processes. The proposed linkage error models are built directly on the data values from the original sources, and based on the actual record linkage method that is used. We have adapted the jackknife methods to estimate bias, variance, and mean squared error of our proposed estimators. To illustrate the general methodology, we give one example of estimating the regression coefficients in the linear and logistic regression models, and another example of estimating small area mean under the nested-error linear regression model. In order to reduce the computational burden, simplified version of the proposed estimators, jackknife methods, and numerical algorithms are given. A Monte Carlo simulation study is devised to evaluate the performance of the proposed estimators and to investigate the difference between the standard and simplified jackknife methods.Item Stochastic Simulation: New Stochastic Approximation Methods and Sensitivity Analyses(2015) Chau, Marie; Fu, Michael C.; Applied Mathematics and Scientific Computation; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)In this dissertation, we propose two new types of stochastic approximation (SA) methods and study the sensitivity of SA and of a stochastic gradient method to various input parameters. First, we summarize the most common stochastic gradient estimation techniques, both direct and indirect, as well as the two classical SA algorithms, Robbins-Monro (RM) and Kiefer-Wolfowitz (KW), followed by some well-known modifications to the step size, output, gradient, and projection operator. Second, we introduce two new stochastic gradient methods in SA for univariate and multivariate stochastic optimization problems. Under a setting where both direct and indirect gradients are available, our new SA algorithms estimate the gradient using a hybrid estimator, which is a convex combination of a symmetric finite difference-type gradient estimate and an average of two associated direct gradient estimates. We derive variance minimizing weights that lead to desirable theoretical properties and prove convergence of the SA algorithms. Next, we study the finite-time performance of the KW algorithm and its sensitivity to the step size parameter, along with two of its adaptive variants, namely Kesten's rule and scale-and-shifted KW (SSKW). We conduct a sensitivity analysis of KW and explore the tightness of an mean-squared error (MSE) bound for quadratic functions, a relevant issue for determining how long to run an SA algorithm. Then, we propose two new adaptive step size sequences inspired by both Kesten's rule and SSKW, which address some of their weaknesses. Instead of us- ing one step size sequence, our adaptive step size is based on two deterministic sequences, and the step size used in the current iteration depends on the perceived proximity of the current iterate to the optimum. In addition, we introduce a method to adaptively adjust the two deterministic sequences. Lastly, we investigate the performance of a modified pathwise gradient estimation method that is applied to financial options with discontinuous payoffs, and in particular, used to estimate the Greeks, which measure the rate of change of (financial) derivative prices with respect to underlying market parameters and are central to financial risk management. The newly proposed kernel estimator relies on a smoothing bandwidth parameter. We explore the accuracy of the Greeks with varying bandwidths and investigate the sensitivity of a proposed iterative scheme that generates an estimate of the optimal bandwidth.Item A Probabilistic Estimation Model for the Recoverable Leakage of the Water Distribution Network(2014) Ghoniema, Moatassem Mohamed; Ayyub, Bilal M; Civil Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Water Distribution Networks (WDN) play a vitally important role in preserving and providing a desirable life quality to the public. A WDN should provide, during its economic life, the required quality and quantity of water at required pressures. Leakage rate and its high associated cost of failure have reached a level that now draws the attention of both policy and decision makers. Leakage is usually the major cause of water loss in water distribution systems. EPA reported in 2007, 240,000 water main breaks per year in the US. The USGS in 2007 estimated that water lost from water distribution systems is 1.7 trillion gallons per year at a national cost of $2.6 billion per year. Leakage occurs in different components of the water distribution system. Causes of leaks include corrosion, soil corrosivity, excessive water pressure, material defects, water hammer, excessive loads and vibration from road traffic and stray electric current. In this dissertation a probabilistic estimation model for the recoverable leakage of WDNs was presented factoring key causes that lead to high percentages of leakage in different components of the WDN. The model receives the deterministic and stochastic description of the leakage of the WDN received from the research survey. It is evident that IWA's model for estimating Unavoidable Annual Real Losses (UARL) does not account for soil corrosivity. The UARL equation can be modified by adding a new soil corrosivity factor (Cr) that takes the soil corrosivity into consideration. Linear Regression was used to develop a relationship between the UARL and the soil corrosivity. Directional cosines analysis examined the importance of the random variables in the new probabilistic estimation model. Two Case Studies were used to validate the modified formulation for the UARL using the data for the leakage component parameters and the system water audit. Monte Carlo simulation was operated twice till the distribution had minimal change. After adding the Cr the output distributions for the UARL had a 43% decrease in the standard deviation value which shows that the corrosion behavior of WDNs is closely related to the environmental factors.Item Simulating and Optimizing: Military Manpower Modeling and Mountain Range Options(2009) Hall, Andrew Oscar; Fu, Michael C; Business and Management: Decision & Information Technologies; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)In this dissertation we employ two different optimization methodologies, dynamic programming and linear programming, and stochastic simulation. The first two essays are drawn from military manpower modeling and the last is an application in finance. First, we investigate two different models to explore the military manpower system. The first model describes the optimal retirement behavior for an Army officer from any point in their career. We address the optimal retirement policies for Army officers, incorporating the current retirement system, pay tables, and Army promotion opportunities. We find that the optimal policy for taste-neutral Lieutenant Colonels is to retire at 20 years. We demonstrate the value and importance of promotion signals regarding the promotion distribution to Colonel. Signaling an increased promotion opportunity from 50% to 75% for the most competitive officers switches their optimal policy at twenty years to continuing to serve and competing for promotion to Colonel. The second essay explores the attainability and sustainability of Army force profiles. We propose a new network structure that incorporates both rank and years in grade to combine cohort, rank, and specialty modeling without falling into the common pitfalls of small cell size and uncontrollable end effects. This is the first implementation of specialty modeling in a manpower model for U.S. Army officers. Previous specialty models of the U.S. Army manpower system have isolated accession planning for Second Lieutenants and the Career Field Designation process for Majors, but this is the first integration of rank and specialty modeling over the entire officer's career and development of an optimal force profile. The last application is drawn from financial engineering and explores several exotic derivatives that are collectively known Mountain Range options, employing Monte Carlo simulation to price these options and developing gradient estimates to study the sensitivities to underlying parameters, known as "the Greeks". We find that IPA and LR/SF methods are efficient methods of gradient estimation for Mountain Range products at a considerably reduced computation cost compared with the commonly used finite difference methods.