UMD Theses and Dissertations
Permanent URI for this collectionhttp://hdl.handle.net/1903/3
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 given thesis/dissertation in DRUM.
More information is available at Theses and Dissertations at University of Maryland Libraries.
Browse
8 results
Search Results
Item Dynamic Traffic Management of Highway Networks(2022) Alimardani, Fatemeh; Baras, John S; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Efficient operation of traffic networks via management strategies can guarantee overall societal benefits for both the humans and the environment. As the number of vehicles and the need for transportation grows, dynamic traffic management aims to increase the safety and efficiency of the traffic networks without the need to change the infrastructure of the existing roads. Since the highway networks are considered permanent investments that are expensive to build and maintain, the main scope of this dissertation is to propose traffic flow models and methods to improve the efficiency of the current highway systems without the need to change their infrastructure. When all vehicles in a network are \textit{Human-Driven Vehicles} (HDVs), and changing the infrastructure is either so expensive or impossible, then one reasonable approach to improve the efficiency of traffic networks is through the control of traffic signal lights specially because the behavior of the human drivers cannot be directly controlled. A literature review of highway traffic control demonstrate that \textit{Ramp Metering} (RM) is one of the most commonly used approaches as it improves the network performance in regards to travel time, travel distance, throughput, etc and cost-wise, it is a very economical approach. As such, in this research, the ultimate goal focus is to extend the current literature on traffic managements of highway networks by offering new models and algorithms to improve this field. To reach this goal, the first step is to focus on improving and extending the current traffic flow models. There are two categories of traffic flow models in the literature: First-order models, and Second-order models. Many different extensions of the famous first-order model called the Cell-Transmission Model (CTM) have been proposed throughout the past decades, each one proposed based on different criteria and the specific needs of different applications. In the first part of this dissertation, a performance assessment of the most important extensions of CTM will be performed. Then, based on this evaluation, an extended version of the CTM, called the Piece-Wise Affine Approximation-CTM (\textit{PWA-CTM}), will be offered which will be proven to have better performance regarding the evolution of traffic flow and computation time comparing to the previous versions of this model. In the next step, the focus will be shifted to second-order models as they have better capabilities of modeling the behavior of traffic flow comparing to the first-order models. However, any optimization scheme for highway traffic control based on these models is highly nonlinear and computationally intensive. As such, in this part of the research, a linearization of the famous second-order model called the \textit{METANET} will be offered which is based on PWA approximations and also synthetic data generation techniques. With extensive simulations, it will be shown that this linearized approximation can greatly impact the computational complexity of any optimization-based traffic control framework based on this second-order traffic flow model. Moreover, to have significant traffic management improvements, not only the underlying traffic models, but also the control strategies should be enhanced. The availability of increasing computational power and sensing and communication capabilities, as well as advances in the field of machine learning, has developed \textit{learning-based} control approaches which can address constraint satisfaction and closed-loop performance optimization. In this chapter, \textit{Reinforcement Learning} (RL) algorithms will be investigated to solve the optimal control problem of RM. In the case of RM, RL-based techniques offer a potentially appealing alternative method to solve the problem at hand, since they are data-based and make no assumptions on the underlying model parameters. Towards this direction, it is convenient to study the road model as a multi-agent system of non-homogeneous networked agents. In the following, a novel formulation of the RM problem as an optimal control problem based on a first-order multi-agent dynamical system will be offered. Then, applying policy gradient RL algorithms, a probabilistic policy will be found that solves the ramp-metering problem. The performance of the optimal policy learnt will be investigated under different scenarios to evaluate its efficiency.Item Vehicle Bridge Interaction analysis on Concrete and Steel Curved Bridges(2017) Ye, Yunchao; Fu, Chung.C; Civil Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)This study investigation is intended to research the dynamic reactions of horizontally curved bridge under heavy vehicle load. Most of the main factors that affect the bridge dynamic response due to moving vehicles are considered. First, an improved grid model is developed for the analysis of curved bridges based on the shear-flexibility grillage analyzing method, in which the effects of warping stiffness and moment of inertia are both considered. Based on commercial software ANSYS Mechanical APDL, 3D beam element, mass element and spring-damper element are integrated together in building a 3D vehicle and bridge system. A simplified numeric method is developed for solving the interaction problem, considering the effect of random road roughness and its velocity term. This system is tested on two case study bridges, Manchuria concrete bridge and Veteran’s Memorial steel bridge. Second, with the model and numerical method presented, a series of parametric studies are conducted to study during the curved bridge dynamic interaction. In the study, the effects of curvature of radius, bridge surface roughness, bridge span configuration, traffic lane eccentricities and speed are investigated and analyzed. The dynamic response and dynamic impact factors are calculated and compared. The analysis results provide good references for the stipulation of impact factor formulae in the later studies. Third, based on the investigation of determining factors of curve bridge dynamic interaction, the expression for upper-bound envelop for impact factors of maximum deflection is given in different surface conditions and highway speed limits as a function of bridge fundamental frequency or bridge central angle. A study is conducted on comparing this empirical equation and serval other major design codes, comments and suggestion are made based on the discoveries.Item UNDERSTANDING AND MODELLING TIME USE, WELL BEING AND DYNAMICS IN ACTIVITY-TRAVEL BEHAVIOR: A CHOICE BASED APPROACH(2018) Dong, Han; Cirillo, Cinzia; Civil Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Understanding the determinants of activity and travel related choices is critical for policy-makers, planners and engineers who are in charge of the management and design of large scale transportation systems. These systems, and their externalities, are interwoven with human actions and communities’ evolution. Traditionally, individual decision-making and travel behaviour studies are based on random utility models (RUM) and discrete choice analysis. To extend the ability of modellers to represent and forecast complex travel behaviour, this dissertation expands existing models to accommodate the influence of variables other than the traditional socio-demographics or level of service variables. In this thesis, technology innovations, psychological factors, and perceptions of future uncertainty are integrated into the classical RUMs and their effects on activity-travel decision making are investigated. Technology innovations, such as telecommunication, online communities and entertainment, release individual’s time and space constraints. They also modify people’s activity and travel choices. An integrated discrete-continuous RUM is proposed to study individuals’ participation in leisure activities, which is an important component of activity scheduling analysis and tour/trip formation. Leisure alternatives considered include: computer/internet related activity, in-home activity, and out-of-home activity. Compared to previous discrete-continuous models, interdependence among activities and the related time usage is explored using a modelling structure that accommodate full correlation among decision variables of different types. Standard random utility models are extended by including attitudes and perceptions as latent variables; these constructs are expected to enhance the behavioural representation of the choice process. A simultaneous structural model is proposed to represent the mutual effects existing between psychological factors and activity choices. Biases due to endogeneity in psychological factors and activity choices are taken into consideration in the model. To further extend the behavioural realism of our model, this thesis proposes a new simultaneous equation model formulation that links psychological indicators to activity participation and time use decisions. Unlike previous studies, the proposed method allows the psychological factors to be correlated with time use decisions and serve as an attribute in time use choice model. A new iterative simulated maximization estimation method is also proposed to accommodate possible endogeneity bias in the model system. A simulation experiment shows that the estimation method produces consistent and unbiased estimation results. Moreover, a real case study is also implemented in the context of participation in leisure activities, linking emotions, activity involvement and time use. After exploring individual’s decisions on activity and time use choices, a dynamic discrete choice model framework is proposed to accommodate stochasticity in individual behaviour over time. Following previous studies, activity patterns are decomposed into tour and stop sequences. Accordingly, a tour choice model and a stop choice model are jointly formulated under a unified framework with a hierarchical structure where stop choices are assumed to be conditional on tour choices. The results indicate that individuals are sensitive to current and future changes in travel and activity characteristics and that a dynamic formulation better represents multi-day travel behaviour.Item ESSAYS ON SKILLS AND VICTIMIZATION(2015) Sarzosa, Miguel; Urzua, Sergio; Economics; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Recent literature has shown that skills are not only essential for the development of successful adults, but also that they are malleable and prone to be affected by many experiences. In this dissertation, I explore these two sides of skills and development. I use skills as a vehicle to study the consequences victimization events have on adult outcomes, and as the channels through which the gaps in those adult outcomes materialize. I use novel longitudinal surveys and rely on an empirical strategy that recognizes skills as latent constructs. First, I estimate the treatment effects being bullied and being a bully in middle school have on several outcomes measured at age 18. I find that both, victims and bullies, have negative consequences later in life. However, they differ in how non-cognitive and cognitive skills palliate or exacerbate these consequences. Then, I move on to investigate the channels that open the gaps in adult outcomes between victims and non-victims. I estimate a structural dynamic model of skill accumulation. I allow skill formation to depend on past levels of skills, parental investment and bullying. Also, I allow bullying itself to depend on each student's past skills and those of his or her classmates. I find that being bullied at age 14 depletes current skill levels by 14% of a standard deviation for the average child. This skill depletion causes the individual to become 25% more likely to experience bullying again at age 15. Therefore bullying triggers a self-reinforcing mechanism that opens an ever-growing skill gap that reaches about one standard deviation by age 16. Finally, under the light of skills, I explore how other type of victimization, namely discrimination against sexual minorities, creates income gaps against non- heterosexual workers. I estimate a structural model that relies on the identification of unobserved skills to allow schooling choices, occupational choices and labor market outcomes to be endogenously determined and affected by the sexual preference of the worker. The results show that difference in skills, observable characteristics, and tastes for tertiary education and type of occupation, contribute to at least half of the income gaps non-heterosexuals face.Item The Dynamics of Political Participation: An Analysis of the Dynamic Interaction between Individuals and their Political Micro-Environment(2012) Wendel, Stephen A.; Oppenheimer, Joe; Government and Politics; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)While political choices are rarely isolated or simultaneous, the vast majority of empirical models in political science assume they are. This dissertation examines the dynamic interactions over time between individuals and their micro-environment, in which a single factor both influences, and is influenced by, the act of voting. These dynamic interactions occur in a surprisingly broad swathe of the current literature on American voting behavior, as implicit but unexamined elements of four major research traditions. When these interactions are present, they establish feedback cycles that pose both theoretical and statistical challenges if not analyzed appropriately. Researchers ignoring these cycles tend to underestimate long term influences on voting behavior, make unrealistic assumptions about changes in voting behavior over time, and produce biased results under certain conditions. I propose a methodology that can successfully identify and model these interactions: employing simulation models to represent dynamic interactions in an intuitive format, and using optimization techniques to conduct parameter estimation and hypothesis testing against empirical data. To guide the development of these simulation models, I outline a theoretical framework of the major pathways by which dynamic interactions can influence voting behavior. I then present two applications of this methodology, to study the dynamic impacts on voting of political mobilization, and of social conformity over time. In both cases, the models receive strong statistical support, in benchmark tests against existing econometric models and against empirical data on voting behavior. Both mobilization and social conformity have unstudied indirect impacts that can lead to an additional 1.7% to 4% increase in voter turnout beyond existing models. Targeted use of peer pressure can lead to even more significant increases in turnout - up to a 30% increase among otherwise indecisive voters. In the long term, targeted mobilization can create cadres of repeatedly-mobilized activists, which raises questions about whether political campaigns effectively use their mobilization funds to build their parties in the long term. These two simulation models also provide a foundation for a host of new research questions, ranging from the impact of high-intensity get-out-the-vote drives on future mobilization efforts, to the effects of an aging population on turnout behavior over time.Item Generic Dynamic Model for a Range of Thermal System Components(2010) Xuan, Shenglan; Radermacher, Reinhard; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)The simulation of a thermal system consists of a simulation of its components and their interactions. The advantages of thermal system simulations have been widely recognized. They can be used to explore the performance of a newly designed system, to identify whether the design meets the design criteria, to develop and test controls, and to optimize the system by minimizing the cost or power consumption, and maximizing the energy efficiency and/or capacity. Thermal system simulations can also be applied to existing systems to explore prospective modifications and improvements. Much research has been conducted on aspects of thermal system and component simulation, especially for steady-state simulation. Recently, transient simulations for systems and components have gained attention, since dynamic modeling assists the understanding of the operation of thermal systems and their controls. This research presents the development of a generic component model that allows users to easily create and customize any thermal component with a choice of working fluids and levels of complexity for either transient or steady-state simulation. The underlying challenge here is to design the code such that a single set of governing equations can be used to accurately describe the behavior of any component of interest. The inherent benefits to this approach are that maintenance of the code is greatly facilitated as compared to competing approaches, and that the software is internally consistent. This generic model features a user-friendly description of component geometry and operating conditions, interactive data input and output, and a robust component solver. The open literature pertaining to thermal component models, especially the components of vapor compression systems, is reviewed and commented on in this research. A theoretical evaluation of the problem formulation and solution methodology is conducted and discussed. A generic structure is proposed and developed to simulate thermal components by enabling and disabling a portion of the set of governing equations. In addition, a system solver is developed to solve a system composed of these components. The component/system model is validated with experimental data, and future work is outlined.Item Expanding on Architecture: A New School of Architecture Planning and Preservation, UMCP(2007-12-17) Talbott, Michael; Williams, Isaac; Architecture; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)This thesis explores the limits of the architectural design process by proposing continuous and evolving vision of space and form as a dynamic and adaptive response to changes in context. The document defines a restructured framework of architecture in time. The theory prescribes a dynamic architecture, able to evolve and transform over the course of its life for the good of ecological and functional sustainability. The result demonstrates the benefits and challenges of a dynamic design process applied to the future expansion of the University of Maryland School of Architecture, Planning and Preservation. This thesis evaluates the current condition of the school, identifies the opportunities and issues, and designs the architectural interventions and additions necessary to satisfy the current and future needs of the school. The result addresses any identified programmatic issues in a series of sequential architectural propositions over the next 8 years. The effort focuses on the following question: How can architecture be designed to better adapt to contextual changes over time to create more efficient, more functional, and more beautiful architecture and that avoids obsolescence and environmental degradation?Item DEVELOPMENT OF A DYNAMIC TEST FACILITY FOR ENVIRONMENTAL CONTROL SYSTEMS(2006-04-06) Gado, Amr El-Sayed Alaa El-Din; Radermacher, Reinhard; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Passenger cars and light trucks consume 80% of the total oil imported by U.S.A. Mobile air conditioners (MACs) increase vehicle fuel consumption and exhaust gas emissions. They operate most of the time in a transient state. It is currently impossible to test the performance of an air conditioner during transient operation without it being associated with its intended conditioned space, the car cabin. In this research work a new smart test facility is designed, built, and verified. This facility makes it possible to test the MAC independent of the vehicle, but yet under realistic dynamic conditions. The facility depends on simulation software that measures the conditions of the air supplied by the MAC and subsequently adjusts the conditions of the air returning to the MAC depending on the results of a thermal numerical model of the car cabin that takes into consideration sensible and latent loads, as well as passengers' control settings. It was successful in controlling the temperature and relative humidity within ±0.9°C and ±5% of their respective intended values. The test facility is used to investigate the dynamic performance of a typical R134a MAC system. The tests include pull-down, drive cycle, and cyclic on/off tests. The analysis focuses on the latent capacity and moisture removal due to the difficulty in measuring these variables during field tests. The results show that the most energy efficient method to pull-down the air temperature inside a hot-soaked cabin is to start with fresh air as long as the temperature in the cabin exceeds that of the ambient and then switch to recirculated air. The effect of re-evaporation is illustrated by showing the off-cycle latent capacity. Cyclic tests show that the net moisture removal rate has a minimum at around a 2 minute duty cycles. This implies a means of controlling the coil latent heat factor by varying duty cycle. The automotive air conditioning system is numerically modeled and used in cooperation with the cabin model to conduct numerical tests. The numerical simulation results are compared to the experimental results and the error is less than 1.5 K of cabin air temperature.