Search results

130,177 results

View results as:
Number of results to display per page
Book
1 online resource.
We introduce a fast solver for sparse matrices arising from the finite element discretization of elliptic partial differential equations (PDEs). We use a fast direct (but approximate) multifrontal solver as a preconditioner, and use an iterative solver to achieve a desired accuracy. This approach combines the advantages of direct and iterative schemes to arrive at a fast, robust and accurate solver. We will show that this solver is much faster and more memory efficient compared to a conventional direct multifrontal solver. Furthermore, we will demonstrate that the solver is both a faster and more effective preconditioner than other preconditioners such as the incomplete LU (ILU) preconditioner. The solver can be applied to both structured and unstructured meshes in a similar manner. We build on our previous work and utilize the fact that dense frontal and update matrices, in the multifrontal algorithm, can be represented as hierarchically off-diagonal low-rank (HODLR) matrices. Using this idea, we replace all large dense matrix operations in the multifrontal elimination process with O(N) HODLR operations to arrive at a faster and more memory efficient solver.
We introduce a fast solver for sparse matrices arising from the finite element discretization of elliptic partial differential equations (PDEs). We use a fast direct (but approximate) multifrontal solver as a preconditioner, and use an iterative solver to achieve a desired accuracy. This approach combines the advantages of direct and iterative schemes to arrive at a fast, robust and accurate solver. We will show that this solver is much faster and more memory efficient compared to a conventional direct multifrontal solver. Furthermore, we will demonstrate that the solver is both a faster and more effective preconditioner than other preconditioners such as the incomplete LU (ILU) preconditioner. The solver can be applied to both structured and unstructured meshes in a similar manner. We build on our previous work and utilize the fact that dense frontal and update matrices, in the multifrontal algorithm, can be represented as hierarchically off-diagonal low-rank (HODLR) matrices. Using this idea, we replace all large dense matrix operations in the multifrontal elimination process with O(N) HODLR operations to arrive at a faster and more memory efficient solver.
Book
1 online resource.
This dissertation examines the creation, demarcation, and enforcement of the Canadian-United States border from the American Revolution until 1924. Although diplomats often viewed the division between national spaces as a clean and uniform line, the Canadian-United States border is better understood as a series of interlocking, uneven, and inconsistent walls. Federal administrators created national border policy, but local agents dictated the practical impacts this policy could have. Inconsistent policy decisions ensured that the border closed at different times in different regions and created an uneven set of impositions on borderland communities. As a result, federal border policy cannot be understood by looking at a single racial, ethnic, or tribal group. Europeans, Chinese, African Americans, Cree, Sioux, Nez Perce, Métis, Stó:lō, Haida, Ojibwe, and Iroquois all experienced the impacts of border closure in different ways and to different extents.
This dissertation examines the creation, demarcation, and enforcement of the Canadian-United States border from the American Revolution until 1924. Although diplomats often viewed the division between national spaces as a clean and uniform line, the Canadian-United States border is better understood as a series of interlocking, uneven, and inconsistent walls. Federal administrators created national border policy, but local agents dictated the practical impacts this policy could have. Inconsistent policy decisions ensured that the border closed at different times in different regions and created an uneven set of impositions on borderland communities. As a result, federal border policy cannot be understood by looking at a single racial, ethnic, or tribal group. Europeans, Chinese, African Americans, Cree, Sioux, Nez Perce, Métis, Stó:lō, Haida, Ojibwe, and Iroquois all experienced the impacts of border closure in different ways and to different extents.
Book
1 online resource.
The enormous size and cost of current state-of-the-art accelerators based upon conventional radio-frequency (RF) technology has spawned a great interest in developing new acceleration concepts that are more compact and economical. Micro-fabricated dielectric laser accelerators (DLAs) are an attractive approach as such structures can support accelerating fields one to two orders of magnitude higher than RF cavity-based accelerators. DLAs use commercial lasers as a power source, which are smaller and less expensive than RF klystrons that power today's accelerators. In addition, DLAs are fabricated via mass-producible, low cost, lithographic techniques. However, despite several DLA structures being proposed recently, no successful demonstration of acceleration in these structures had been shown until this work. This thesis reports the first observation of high-gradient (exceeding 300 MeV/m) acceleration of electrons in a DLA. Relativistic (60 MeV) electrons are energy modulated over 563 optical periods of a fused silica grating structure, powered by a 800 nm wavelength mode-locked Ti:Sapphire laser. The observed results are in agreement with analytical models and electrodynamic simulations. By comparison, conventional modern linear accelerators operate at gradients of 10-30 MeV/m; and the first linear RF cavity accelerator was 10 RF periods (1 m long) with a gradient of approximately 1.6 MV/m. Our results set the stage for the development of future multi-staged DLA devices composed of integrated on-chip systems. This would enable compact table-top MeV to GeV scale accelerators for security scanners and medical therapy, university-scale x-ray light sources for biological and materials research, portable medical imaging devices, and would substantially reduce the size and cost of a future multi-TeV scale collider.
The enormous size and cost of current state-of-the-art accelerators based upon conventional radio-frequency (RF) technology has spawned a great interest in developing new acceleration concepts that are more compact and economical. Micro-fabricated dielectric laser accelerators (DLAs) are an attractive approach as such structures can support accelerating fields one to two orders of magnitude higher than RF cavity-based accelerators. DLAs use commercial lasers as a power source, which are smaller and less expensive than RF klystrons that power today's accelerators. In addition, DLAs are fabricated via mass-producible, low cost, lithographic techniques. However, despite several DLA structures being proposed recently, no successful demonstration of acceleration in these structures had been shown until this work. This thesis reports the first observation of high-gradient (exceeding 300 MeV/m) acceleration of electrons in a DLA. Relativistic (60 MeV) electrons are energy modulated over 563 optical periods of a fused silica grating structure, powered by a 800 nm wavelength mode-locked Ti:Sapphire laser. The observed results are in agreement with analytical models and electrodynamic simulations. By comparison, conventional modern linear accelerators operate at gradients of 10-30 MeV/m; and the first linear RF cavity accelerator was 10 RF periods (1 m long) with a gradient of approximately 1.6 MV/m. Our results set the stage for the development of future multi-staged DLA devices composed of integrated on-chip systems. This would enable compact table-top MeV to GeV scale accelerators for security scanners and medical therapy, university-scale x-ray light sources for biological and materials research, portable medical imaging devices, and would substantially reduce the size and cost of a future multi-TeV scale collider.
Book
1 online resource.
Dynamic state-space models are useful for describing data in various fields, including robotics. An important problem that may be solved by using dynamic state-space models is the estimation of underlying state processes from given observations. When the models are non-linear and the noise not Gaussian, it is impossible to solve the problem analytically; thus, particle filters, also known as sequential Monte Carlo methods, tend to be employed. However, because particle filters are based on sequential importance sampling, the problem arises of how to select the importance density function. Handling unknown parameters in the model presents another significant difficulty in particle filtering. Simultaneous localization and mapping (SLAM) in robotics is one well-known but difficult problem for which particle filters have been used. This dissertation is motivated by SLAM problems and related particle filtering approaches. In this dissertation, we design a new proposal distribution that better approximates the optimal importance function, using a novel way of combining information from observations and state transition dynamics. In the first part of our study, after reviewing representative approaches for SLAM problems, we justify our method of combining information with a series of examples and offer an efficient means of constructing the new proposal distribution. In the second part, we focus on the problems inherent in handling unknown parameters in state-space models. We suggest the application of one-step recursive expectation-maximization (EM) algorithm to learn unknown parameters, and recommend pairing it with the new proposal distribution into an adaptive particle filter algorithm. Furthermore, we propose a new SLAM filter based on the adaptation of the new adaptive particle filter to SLAM problems. In Chapter 3, we conduct simulation studies on localization and SLAM problems to demonstrate the superior numerical performance of the proposed algorithms.
Dynamic state-space models are useful for describing data in various fields, including robotics. An important problem that may be solved by using dynamic state-space models is the estimation of underlying state processes from given observations. When the models are non-linear and the noise not Gaussian, it is impossible to solve the problem analytically; thus, particle filters, also known as sequential Monte Carlo methods, tend to be employed. However, because particle filters are based on sequential importance sampling, the problem arises of how to select the importance density function. Handling unknown parameters in the model presents another significant difficulty in particle filtering. Simultaneous localization and mapping (SLAM) in robotics is one well-known but difficult problem for which particle filters have been used. This dissertation is motivated by SLAM problems and related particle filtering approaches. In this dissertation, we design a new proposal distribution that better approximates the optimal importance function, using a novel way of combining information from observations and state transition dynamics. In the first part of our study, after reviewing representative approaches for SLAM problems, we justify our method of combining information with a series of examples and offer an efficient means of constructing the new proposal distribution. In the second part, we focus on the problems inherent in handling unknown parameters in state-space models. We suggest the application of one-step recursive expectation-maximization (EM) algorithm to learn unknown parameters, and recommend pairing it with the new proposal distribution into an adaptive particle filter algorithm. Furthermore, we propose a new SLAM filter based on the adaptation of the new adaptive particle filter to SLAM problems. In Chapter 3, we conduct simulation studies on localization and SLAM problems to demonstrate the superior numerical performance of the proposed algorithms.
Book
727 pages : illustrations ; 24 cm.
Green Library
Status of items at Green Library
Green Library Status
Stacks Find it
DD491 .H2 H5 V.279 Unavailable In process Request
Book
1 online resource.
Symmetric cone optimization subsumes linear optimization, second-order cone optimization, and semidefinite optimization. It is of interest to extend the algorithmic developments of symmetric cone optimization into the realm of unsymmetric cones. We analyze the theoretical properties of some algorithms for unsymmetric cone problems. We show that they achieve excellent worst-case iteration bounds while not necessarily being practical to implement. Using lessons from this analysis and inspired by the Mehrotra predictor-corrector algorithm, we extend the homogeneous implementation ECOS to handle problems modeled with Cartesian products of the positive orthant, second-order cones, and the exponential cone, and we empirically validate its efficiency.
Symmetric cone optimization subsumes linear optimization, second-order cone optimization, and semidefinite optimization. It is of interest to extend the algorithmic developments of symmetric cone optimization into the realm of unsymmetric cones. We analyze the theoretical properties of some algorithms for unsymmetric cone problems. We show that they achieve excellent worst-case iteration bounds while not necessarily being practical to implement. Using lessons from this analysis and inspired by the Mehrotra predictor-corrector algorithm, we extend the homogeneous implementation ECOS to handle problems modeled with Cartesian products of the positive orthant, second-order cones, and the exponential cone, and we empirically validate its efficiency.
Book
1 online resource.
While fossil fuels have helped our society grow and technology advance rapidly in the last century, the mounting evidence of their detrimental effects on air quality, public health, and climate change, in addition to the desire for energy independence from geopolitically instable suppliers, are strong motivations for shifting to cleaner, independent energy sources. Wind and solar are promising alternatives, but their intermittency leads to either curtailing or the need for energy storage. One approach is the carbon dioxide reduction reaction (CO2RR), which combines CO2 and H2O using renewable energy to produce high energy density fuels and commodity chemicals, providing an attractive, lower-carbon alternative process to fossil fuels and petrochemicals. However, a wide variety of products can be produced from this reaction, and there is still much work to be done in developing catalysts to steer the selectivity, increase activity, and maximize energy efficiency for the CO2RR in order for the process to be cost-competitive with other solutions. Transition metal surfaces have been explored as electrocatalysts for the CO2RR, with the activity of several metals reported in seminal work by Yoshio Hori. The Jaramillo laboratory designed a new electrolysis cell with excellent product detection sensitivity that not only was able to reproduce Hori's results, but also was able to detect additional products. Additionally, it was used to explore a wider set of potentials to obtain a more complete view of the catalytic activity of several transition metal catalysts. The results of several metals - Pt, Fe, Ni, Cu, Au, Ag, and Zn - were collectively analyzed to provide new insights into the CO2RR on metals, particular with the selectivity of minor products such as methanol and methane. Volcano plots based on theoretically predicted CO binding energy were constructed, and our experimental results correlate well with theoretical predictions. After establishing a reliable testing protocol for the CO2RR, we have since moved to exploring novel materials. Amines are commonly used as CO2 capture sorbents and adsorbents for flue gases of coal-fired power plants, as amines can interact with the CO2 molecules to form carbamates through N-C bond formation, and the CO2 can be released through a temperature or pressure change to produce pure CO2 streams. Additionally, pyridine has been added to solution to modify the activity of metal and semiconductor catalyst surfaces to produce methanol at fairly high Faradaic efficiencies, and the formation of a weak N-C bond has been detected and proposed as a mechanism for the enhanced CO2RR activity. Ionic liquids containing amine groups have also shown interesting activity for the CO2RR, and these works inspired our work to use a thin film amine- containing polymer as a surface modifier to affect the activity of metal catalysts. Specifically, thin films of 10-20nm of polyaniline (PAni) were electrodeposited on a polycrystalline Pt foil and explored for the CO2RR. Up to a 5x enhancement in the production of formate as well as an increase in the production of CO at high overpotentials were observed for the PAni-Pt films relative to the Pt foils, and these enhancements were confirmed to not be due to PAni degradation. These results show a promising method for modifying CO2RR activity that could be used and optimized for other similar systems in the future. Alloys have been interesting systems to explore for a wide variety of electrocatalytic reactions, and high-throughput testing as well as theoretical predictions are two methods to try and reduce the time needed to find active catalysts. Due to the need for product detection, it is difficult to design a high-throughput CO2RR system, and the complexity of the reaction also makes it difficult to find theoretical studies that have accurately predicted active catalysts to this point. Recent theoretical studies have provided some intriguing candidates by simplifying the reaction to just a few rate- determining steps, and some progress has been made in correlating theoretical studies with experimental results. We have developed a "medium throughput" testing protocol to be able to synthesize, physically characterize, and electrochemically test about 4 alloy compositions each day and have worked with Jens Norskov's group to explore potential alloy candidates for the CO2RR. A dual e-beam system was designed by my colleague, Christopher Hahn, that can coevaporate a variety of alloy thin films of desired compositions. Code has been written to quickly analyze individual materials and compare results with a growing database of materials to improve the speed of analysis and assessment for trying to narrow down candidates and find interesting CO2RR alloy catalysts. One of the most promising results from the alloy screening to this point has been the PtInx system of alloys. While Pt is known to bind CO strongly and produce primarily H2 and In is known to bind CO weakly and produce primarily formate, several synthesized PtInx alloys predominantly produced CO. These films were then further characterized to gain a more complete understanding of the alloy compositions, crystallinity, and morphology of the alloys, and the results were then analyzed based on the knowledge gained of the catalytic surfaces. The most active alloys had bulk compositions of PtIn4 and PtIn12 , but the surfaces were mostly a crystalline Pt3In7 phase, with phase-separated In particles on the surface as well. It is postulated that the increased activity for CO was due to the Pt3In7 alloy phase, which has a binding energy for CO that is in between that for Pt or In. This was a promising initial results that shows that the activity of alloys can be very different than that for either of the individual elements, and our system will hopefully continue to find additional active alloys and provide insight into the mechanisms for the CO2RR.
While fossil fuels have helped our society grow and technology advance rapidly in the last century, the mounting evidence of their detrimental effects on air quality, public health, and climate change, in addition to the desire for energy independence from geopolitically instable suppliers, are strong motivations for shifting to cleaner, independent energy sources. Wind and solar are promising alternatives, but their intermittency leads to either curtailing or the need for energy storage. One approach is the carbon dioxide reduction reaction (CO2RR), which combines CO2 and H2O using renewable energy to produce high energy density fuels and commodity chemicals, providing an attractive, lower-carbon alternative process to fossil fuels and petrochemicals. However, a wide variety of products can be produced from this reaction, and there is still much work to be done in developing catalysts to steer the selectivity, increase activity, and maximize energy efficiency for the CO2RR in order for the process to be cost-competitive with other solutions. Transition metal surfaces have been explored as electrocatalysts for the CO2RR, with the activity of several metals reported in seminal work by Yoshio Hori. The Jaramillo laboratory designed a new electrolysis cell with excellent product detection sensitivity that not only was able to reproduce Hori's results, but also was able to detect additional products. Additionally, it was used to explore a wider set of potentials to obtain a more complete view of the catalytic activity of several transition metal catalysts. The results of several metals - Pt, Fe, Ni, Cu, Au, Ag, and Zn - were collectively analyzed to provide new insights into the CO2RR on metals, particular with the selectivity of minor products such as methanol and methane. Volcano plots based on theoretically predicted CO binding energy were constructed, and our experimental results correlate well with theoretical predictions. After establishing a reliable testing protocol for the CO2RR, we have since moved to exploring novel materials. Amines are commonly used as CO2 capture sorbents and adsorbents for flue gases of coal-fired power plants, as amines can interact with the CO2 molecules to form carbamates through N-C bond formation, and the CO2 can be released through a temperature or pressure change to produce pure CO2 streams. Additionally, pyridine has been added to solution to modify the activity of metal and semiconductor catalyst surfaces to produce methanol at fairly high Faradaic efficiencies, and the formation of a weak N-C bond has been detected and proposed as a mechanism for the enhanced CO2RR activity. Ionic liquids containing amine groups have also shown interesting activity for the CO2RR, and these works inspired our work to use a thin film amine- containing polymer as a surface modifier to affect the activity of metal catalysts. Specifically, thin films of 10-20nm of polyaniline (PAni) were electrodeposited on a polycrystalline Pt foil and explored for the CO2RR. Up to a 5x enhancement in the production of formate as well as an increase in the production of CO at high overpotentials were observed for the PAni-Pt films relative to the Pt foils, and these enhancements were confirmed to not be due to PAni degradation. These results show a promising method for modifying CO2RR activity that could be used and optimized for other similar systems in the future. Alloys have been interesting systems to explore for a wide variety of electrocatalytic reactions, and high-throughput testing as well as theoretical predictions are two methods to try and reduce the time needed to find active catalysts. Due to the need for product detection, it is difficult to design a high-throughput CO2RR system, and the complexity of the reaction also makes it difficult to find theoretical studies that have accurately predicted active catalysts to this point. Recent theoretical studies have provided some intriguing candidates by simplifying the reaction to just a few rate- determining steps, and some progress has been made in correlating theoretical studies with experimental results. We have developed a "medium throughput" testing protocol to be able to synthesize, physically characterize, and electrochemically test about 4 alloy compositions each day and have worked with Jens Norskov's group to explore potential alloy candidates for the CO2RR. A dual e-beam system was designed by my colleague, Christopher Hahn, that can coevaporate a variety of alloy thin films of desired compositions. Code has been written to quickly analyze individual materials and compare results with a growing database of materials to improve the speed of analysis and assessment for trying to narrow down candidates and find interesting CO2RR alloy catalysts. One of the most promising results from the alloy screening to this point has been the PtInx system of alloys. While Pt is known to bind CO strongly and produce primarily H2 and In is known to bind CO weakly and produce primarily formate, several synthesized PtInx alloys predominantly produced CO. These films were then further characterized to gain a more complete understanding of the alloy compositions, crystallinity, and morphology of the alloys, and the results were then analyzed based on the knowledge gained of the catalytic surfaces. The most active alloys had bulk compositions of PtIn4 and PtIn12 , but the surfaces were mostly a crystalline Pt3In7 phase, with phase-separated In particles on the surface as well. It is postulated that the increased activity for CO was due to the Pt3In7 alloy phase, which has a binding energy for CO that is in between that for Pt or In. This was a promising initial results that shows that the activity of alloys can be very different than that for either of the individual elements, and our system will hopefully continue to find additional active alloys and provide insight into the mechanisms for the CO2RR.
Book
1 online resource.
The adhesive pads on gecko toes are complex systems containing structures at different size scales. Each toe is covered in flaps of skin called lamellae, which are in turn covered in arrays of microscopic hair-like structures known as setae. The tip of each seta splits into hundreds of even smaller nanoscale structures (spatulae) which produce adhesion through intermolecular van der Waals forces. Using this adhesive system, geckos can stick to a wide range of surfaces. One of the most interesting properties of gecko adhesive is controllable adhesion. An adhesive is called controllable if the stickiness can be switched on or off so it can be easily and repeatedly attached and detached. In gecko adhesive, the adhesion is controlled by the shear force: geckos can control their adhesive simply by applying a downwards shear force to their toes. In previous work, a controllable synthetic adhesive was developed that used shear force to control the adhesion similarly to gecko adhesive. The synthetic adhesive consisted of wedge-shaped microstructures made of polydimethylsiloxane (PDMS) silicone rubber, known as microwedges. This thesis presents a new micromachining manufacturing process for microwedge adhesives, which produces stronger adhesives with more varied geometries, enabling practical applications such as grasping and climbing devices for robots and humans. In addition, this thesis investigates the distribution of adhesive stress in natural gecko adhesive and synthetic microwedge adhesive through a combination of experimental measurements and theoretical modeling. In order for an adhesive system to produce the maximum possible adhesive force, the force must be uniformly distributed over the adhesive area. However, until now it was unknown how forces are distributed in gecko adhesive. To address this question and gain understanding of the gecko's adhesive system, the stress distribution over the toes of a live tokay gecko (Gekko gecko) was measured using a custom optical tactile sensor with 100 micrometer spatial resolution based on frustrated total internal reflection (FTIR). Additionally, the stress distribution in the synthetic microwedge adhesive is investigated with a theoretical model that describes the elastic deformation and adhesive interactions of adhesive microstructures. Adhesion is modeled using a cohesive zone model, where the normal and tangential forces generated along the side of the microwedge depend on the separation distance between the microwedge and the surface. Deformation is modeled using a geometrically exact beam model, where the microwedge is treated as a tapered beam undergoing bending, axial, and shear deformation. This modeling approach accurately reproduces the limit curve in force space of microwedge adhesive, describing the relationship between normal and shear force that gives rise to controllable adhesion. In both the tokay gecko toe and the synthetic adhesive, the stress distributions were found to be nonuniform. In the gecko, the normal stress varied significantly at the lamella scale, with compressive stresses observed in some areas even though the net stress over the toe was tensile. Likewise, the model predicts that the normal stress on an adhesive microwedge varies from tensile to compressive along the adhesive interface, with a net stress that is several times smaller than the maximum stress. If the stresses were distributed uniformly, both systems would be capable of supporting much larger loads (around 20 times larger for tokay gecko toes and 5 times larger for microwedges). The proposed model may be useful in evaluating new microwedge structures with modified geometry in order to design a structure that distributes stress more uniformly. Along with the capabilities of the new micromachining process, this could lead to the development of stronger controllable adhesives.
The adhesive pads on gecko toes are complex systems containing structures at different size scales. Each toe is covered in flaps of skin called lamellae, which are in turn covered in arrays of microscopic hair-like structures known as setae. The tip of each seta splits into hundreds of even smaller nanoscale structures (spatulae) which produce adhesion through intermolecular van der Waals forces. Using this adhesive system, geckos can stick to a wide range of surfaces. One of the most interesting properties of gecko adhesive is controllable adhesion. An adhesive is called controllable if the stickiness can be switched on or off so it can be easily and repeatedly attached and detached. In gecko adhesive, the adhesion is controlled by the shear force: geckos can control their adhesive simply by applying a downwards shear force to their toes. In previous work, a controllable synthetic adhesive was developed that used shear force to control the adhesion similarly to gecko adhesive. The synthetic adhesive consisted of wedge-shaped microstructures made of polydimethylsiloxane (PDMS) silicone rubber, known as microwedges. This thesis presents a new micromachining manufacturing process for microwedge adhesives, which produces stronger adhesives with more varied geometries, enabling practical applications such as grasping and climbing devices for robots and humans. In addition, this thesis investigates the distribution of adhesive stress in natural gecko adhesive and synthetic microwedge adhesive through a combination of experimental measurements and theoretical modeling. In order for an adhesive system to produce the maximum possible adhesive force, the force must be uniformly distributed over the adhesive area. However, until now it was unknown how forces are distributed in gecko adhesive. To address this question and gain understanding of the gecko's adhesive system, the stress distribution over the toes of a live tokay gecko (Gekko gecko) was measured using a custom optical tactile sensor with 100 micrometer spatial resolution based on frustrated total internal reflection (FTIR). Additionally, the stress distribution in the synthetic microwedge adhesive is investigated with a theoretical model that describes the elastic deformation and adhesive interactions of adhesive microstructures. Adhesion is modeled using a cohesive zone model, where the normal and tangential forces generated along the side of the microwedge depend on the separation distance between the microwedge and the surface. Deformation is modeled using a geometrically exact beam model, where the microwedge is treated as a tapered beam undergoing bending, axial, and shear deformation. This modeling approach accurately reproduces the limit curve in force space of microwedge adhesive, describing the relationship between normal and shear force that gives rise to controllable adhesion. In both the tokay gecko toe and the synthetic adhesive, the stress distributions were found to be nonuniform. In the gecko, the normal stress varied significantly at the lamella scale, with compressive stresses observed in some areas even though the net stress over the toe was tensile. Likewise, the model predicts that the normal stress on an adhesive microwedge varies from tensile to compressive along the adhesive interface, with a net stress that is several times smaller than the maximum stress. If the stresses were distributed uniformly, both systems would be capable of supporting much larger loads (around 20 times larger for tokay gecko toes and 5 times larger for microwedges). The proposed model may be useful in evaluating new microwedge structures with modified geometry in order to design a structure that distributes stress more uniformly. Along with the capabilities of the new micromachining process, this could lead to the development of stronger controllable adhesives.
Book
1 online resource.
The study of genomic variation within human populations is critical for elucidating the genetic factors that contribute to disease. Identifying and characterizing the genetic architecture of disease advances clinical care by facilitating the development of novel diagnostic tools, the identification of new therapeutic targets, and the practice of personalized treatment for genetic syndromes. The massive volume of genetic data generated by modern genotyping technologies, combined with the informatics challenges of filtering and interpreting these noisy measurements, represent significant obstacles to genomic research. These technical issues necessitate the development of computationally efficient methodologies that leverage raw genotype data for the comparative genomic analysis of complex phenotypes across human subpopulations. In this dissertation, I describe my contributions towards the biomedical study of genetic syndromes using high-throughput genotyping technologies. First, I discuss methods for studying the genome evolution of pre-malignant cancer lesions during progression to breast cancer. Second, I describe algorithms for performing highly accurate variant validation in genomic studies using next generation sequencing. Finally, I present methods for identifying novel disease susceptibility loci in complex diseases using identity by descent mapping in large case-control cohorts.
The study of genomic variation within human populations is critical for elucidating the genetic factors that contribute to disease. Identifying and characterizing the genetic architecture of disease advances clinical care by facilitating the development of novel diagnostic tools, the identification of new therapeutic targets, and the practice of personalized treatment for genetic syndromes. The massive volume of genetic data generated by modern genotyping technologies, combined with the informatics challenges of filtering and interpreting these noisy measurements, represent significant obstacles to genomic research. These technical issues necessitate the development of computationally efficient methodologies that leverage raw genotype data for the comparative genomic analysis of complex phenotypes across human subpopulations. In this dissertation, I describe my contributions towards the biomedical study of genetic syndromes using high-throughput genotyping technologies. First, I discuss methods for studying the genome evolution of pre-malignant cancer lesions during progression to breast cancer. Second, I describe algorithms for performing highly accurate variant validation in genomic studies using next generation sequencing. Finally, I present methods for identifying novel disease susceptibility loci in complex diseases using identity by descent mapping in large case-control cohorts.
Book
1 online resource.
In the age of computers, there has been growing interest in many areas of science and engineering to use numerical techniques to explore problems that are mathematically or analytically intractable. In photonics specifically, there is a desire to design and realize non-periodic nanophotonic structures that outperform their periodic counterparts. One major obstacle to this goal is the necessity to solve the computationally expensive Maxwell's equations repeatedly over a vast parameter space. Here, we will apply the technique of numerical optimization to design aperiodic plasmonic nanostructures. Plasmonics, the study of driven collective oscillations of free electrons in metals, provides for optical modes with subwavelength footprints and is a promising compromise between the small world of electronics and the fast world of photonics. It is well known that nanostructured grooves and slits on a metallic surface are very strong generators and scatterers of surface plasmons, yet the geometries involved make exact analytical solutions impossible. The first step toward designing aperiodic plasmonic groove structures is to identify the fundamental building block of such structures, the metal-air groove, and simplify its mathematical representation so that calculations can be performed very quickly. After the building blocks are described in sufficient detail, the next step is the mathematical model that relates the individual building blocks together in a complete structure. We develop a computation strategy based on a transfer matrix model that can be used to calculate the plasmonic scattering properties of a number of closely spaced grooves to acceptable accuracy with extreme speed. Having characterized the building blocks and the model that glues them together, we apply these results, together with an optimization procedure, to design interesting aperiodic plasmonic groove structures. We first demonstrate a unidirectional launcher of surface plasmons from normally incident light with an extinction ratio of 55 to 1 using only five grooves. The unidirectional launcher is a great starting point due to its simplicity -- relatively few grooves are needed and there is only one operational wavelength, and yet the parameter space is already large enough to necessitate numerical optimization. Building on top of these results, we will conclude with an investigation of structures that split light into specific regions depending on their wavelength. This thesis demonstrates the ability to control the scattering and localization of plasmons using very few carefully chosen scattering elements, and the work presented here can be harnessed to design the next generation of subwavelength photonic devices.
In the age of computers, there has been growing interest in many areas of science and engineering to use numerical techniques to explore problems that are mathematically or analytically intractable. In photonics specifically, there is a desire to design and realize non-periodic nanophotonic structures that outperform their periodic counterparts. One major obstacle to this goal is the necessity to solve the computationally expensive Maxwell's equations repeatedly over a vast parameter space. Here, we will apply the technique of numerical optimization to design aperiodic plasmonic nanostructures. Plasmonics, the study of driven collective oscillations of free electrons in metals, provides for optical modes with subwavelength footprints and is a promising compromise between the small world of electronics and the fast world of photonics. It is well known that nanostructured grooves and slits on a metallic surface are very strong generators and scatterers of surface plasmons, yet the geometries involved make exact analytical solutions impossible. The first step toward designing aperiodic plasmonic groove structures is to identify the fundamental building block of such structures, the metal-air groove, and simplify its mathematical representation so that calculations can be performed very quickly. After the building blocks are described in sufficient detail, the next step is the mathematical model that relates the individual building blocks together in a complete structure. We develop a computation strategy based on a transfer matrix model that can be used to calculate the plasmonic scattering properties of a number of closely spaced grooves to acceptable accuracy with extreme speed. Having characterized the building blocks and the model that glues them together, we apply these results, together with an optimization procedure, to design interesting aperiodic plasmonic groove structures. We first demonstrate a unidirectional launcher of surface plasmons from normally incident light with an extinction ratio of 55 to 1 using only five grooves. The unidirectional launcher is a great starting point due to its simplicity -- relatively few grooves are needed and there is only one operational wavelength, and yet the parameter space is already large enough to necessitate numerical optimization. Building on top of these results, we will conclude with an investigation of structures that split light into specific regions depending on their wavelength. This thesis demonstrates the ability to control the scattering and localization of plasmons using very few carefully chosen scattering elements, and the work presented here can be harnessed to design the next generation of subwavelength photonic devices.
Book
1 online resource.
Aristotle's understanding of the moral psychology and decision-making of the morally good person is more intellectualist than commentators tend to think. The morally good person must have a complete understanding of the value and structure of virtue. Drawing on a number of under-studied texts and concepts in Aristotle's corpus, I argue that this intellectualism lies at the heart of his ethical theory and cannot be excised with ease. Recognizing the centrality of moral intellectualism to Aristotle's theory of practical wisdom has direct consequences for contemporary theorists who wish to defend an Aristotelian theory of ethics or civic education.
Aristotle's understanding of the moral psychology and decision-making of the morally good person is more intellectualist than commentators tend to think. The morally good person must have a complete understanding of the value and structure of virtue. Drawing on a number of under-studied texts and concepts in Aristotle's corpus, I argue that this intellectualism lies at the heart of his ethical theory and cannot be excised with ease. Recognizing the centrality of moral intellectualism to Aristotle's theory of practical wisdom has direct consequences for contemporary theorists who wish to defend an Aristotelian theory of ethics or civic education.
Book
1 online resource.
This study examines the role played by the set of ascetic precepts known as the dhutagunas in early Indian Buddhist monastic communities. Although there are numerous references to the dhutagunas found throughout the monastic discourses which comprise the Buddhist Vinaya (law codes), modern scholarship has largely rejected the notion that the dhutagunas could have been regarded by members of the monastic community as anything other than vestigial ideals received from the non-Buddhist ascetic milieu. This dissertation challenges the premise that the dhutagunas were merely vestiges from non-Buddhist communities, arguing instead that the editors of the Vinaya viewed the ascetic precepts as practices common to the lifestyle of the early Buddhist monk.
This study examines the role played by the set of ascetic precepts known as the dhutagunas in early Indian Buddhist monastic communities. Although there are numerous references to the dhutagunas found throughout the monastic discourses which comprise the Buddhist Vinaya (law codes), modern scholarship has largely rejected the notion that the dhutagunas could have been regarded by members of the monastic community as anything other than vestigial ideals received from the non-Buddhist ascetic milieu. This dissertation challenges the premise that the dhutagunas were merely vestiges from non-Buddhist communities, arguing instead that the editors of the Vinaya viewed the ascetic precepts as practices common to the lifestyle of the early Buddhist monk.
Book
1 online resource.
Due to its intermittent nature, large-scale adoption of solar energy requires new technological advancements to efficiently store and distribute energy. The photoelectrochemical (PEC) splitting of water is a promising way to capture solar energy and store it in the form of chemical bonds. We look at leveraging the advantages of ALD, a technique well known in the microelectronics industry, to address some of the most pressing issues in PEC water splitting. In particular, the focus of our studies is the development of catalysts to drive the oxygen evolution reaction (OER), a reaction typically associated with high overpotentials and sluggish kinetics. We first investigate known active transition metal oxide catalysts, exploring how to enhance their activity with higher surface area and through electronic effects. We create highly active electrocatalysts of both MnOx and NiOx, and discuss some of the advantages and limitations of using ALD to deposit these films. Next, we focus on using ALD to manage charge transport limitations in semiconducting oxide thin films. We demonstrate the sensitivity of semiconducing thin films to film thickness using ALD TiO2 as a model material. We then show how ALD can be used to explore new semiconducting oxide catalysts, focusing on a Ti-Mn oxide system. We also discuss the integration of these catalysts into PEC devices, with an emphasis on the role of stability, oxidation, and surface area in enhancing the OER activity for photoanodes.
Due to its intermittent nature, large-scale adoption of solar energy requires new technological advancements to efficiently store and distribute energy. The photoelectrochemical (PEC) splitting of water is a promising way to capture solar energy and store it in the form of chemical bonds. We look at leveraging the advantages of ALD, a technique well known in the microelectronics industry, to address some of the most pressing issues in PEC water splitting. In particular, the focus of our studies is the development of catalysts to drive the oxygen evolution reaction (OER), a reaction typically associated with high overpotentials and sluggish kinetics. We first investigate known active transition metal oxide catalysts, exploring how to enhance their activity with higher surface area and through electronic effects. We create highly active electrocatalysts of both MnOx and NiOx, and discuss some of the advantages and limitations of using ALD to deposit these films. Next, we focus on using ALD to manage charge transport limitations in semiconducting oxide thin films. We demonstrate the sensitivity of semiconducing thin films to film thickness using ALD TiO2 as a model material. We then show how ALD can be used to explore new semiconducting oxide catalysts, focusing on a Ti-Mn oxide system. We also discuss the integration of these catalysts into PEC devices, with an emphasis on the role of stability, oxidation, and surface area in enhancing the OER activity for photoanodes.
Book
xv, 337 pages : illustrations ; 24 cm.
Green Library
Status of items at Green Library
Green Library Status
Stacks Find it
(no call number) Unavailable On order Request
Book
1 online resource.
The principal purpose of this work is to develop an autonomous and unsupervised learning methodology for control problems that have continuous state/action spaces. The purpose of this learning is to build a control policy that generally works well for random initial states and achieves a locally minimal cost for a selected initial state. As a performance index, the discounted quadratic cost function was used. In this work, we present a learning method originally designed for continuous state/action space control problems. The proposed learning method has two learning stages, what we call phase-I learning and phase-II learning. Phase-I learning is a learning process that builds a working policy for random initial states. Phase-I learning starts from a randomly initialized policy, and the learning agent gradually builds up control knowledge by adding locally optimal actions, calculated based on the local dynamics estimation and the Riccati recursion, to the policy. The resulting policy can control the system for most of the random initial states. The second state, phase-II learning, is a policy refinement process for a given initial state, and it can be understood as a trajectory optimization process. For phase-II learning, a gradient descent trajectory optimization algorithm is developed. Using the trajectory optimization algorithm, the learning agent gradually adjusts the trajectory, which corresponds to the given initial state, to incrementally improve the policy. Phase-II learning continues until the cost reaches its local minimum; the resulting policy is a locally optimal policy for the given initial state. To test our method, we used three systems, the double integrator with gravity (2D), the single inverted pendulum (4D), and the double inverted pendulum on a card (6D). The results show that the proposed learning can build locally optimal policies within a reasonable number of learning cycles (less than 300 cycles in our experiments). The results also show that only a small portion of the state space was visited during the learning process, which supports the sparsity of the policy approximator grid (PAG). The main contributions of this work are as follows: 1) Development of an autonomous and unsupervised learning method that is designed for and effectively works for continuous state/action space control problems. 2) Development of data structures, which we call the policy approximator grid (PAG), the transition memory set (TMSET), and the action memory set (AMSET), that work seamlessly together with the proposed learning method. The insertion and retrieval times of these data structures are constant and independent of the amount of data that the data structures contain. 3) Experimental results that show that only a small portion of the state space is actually visited during the learning, as the state dimension increases. This "sparsity" allows the learning agent to govern the learning process without an exponentially increasing memory usage for the policy approximator implementation. 4) Development of a gradient descent trajectory optimization algorithm for phase-II learning. This algorithm is a general form of the 1st-order differential dynamic programming (DDP) algorithm.
The principal purpose of this work is to develop an autonomous and unsupervised learning methodology for control problems that have continuous state/action spaces. The purpose of this learning is to build a control policy that generally works well for random initial states and achieves a locally minimal cost for a selected initial state. As a performance index, the discounted quadratic cost function was used. In this work, we present a learning method originally designed for continuous state/action space control problems. The proposed learning method has two learning stages, what we call phase-I learning and phase-II learning. Phase-I learning is a learning process that builds a working policy for random initial states. Phase-I learning starts from a randomly initialized policy, and the learning agent gradually builds up control knowledge by adding locally optimal actions, calculated based on the local dynamics estimation and the Riccati recursion, to the policy. The resulting policy can control the system for most of the random initial states. The second state, phase-II learning, is a policy refinement process for a given initial state, and it can be understood as a trajectory optimization process. For phase-II learning, a gradient descent trajectory optimization algorithm is developed. Using the trajectory optimization algorithm, the learning agent gradually adjusts the trajectory, which corresponds to the given initial state, to incrementally improve the policy. Phase-II learning continues until the cost reaches its local minimum; the resulting policy is a locally optimal policy for the given initial state. To test our method, we used three systems, the double integrator with gravity (2D), the single inverted pendulum (4D), and the double inverted pendulum on a card (6D). The results show that the proposed learning can build locally optimal policies within a reasonable number of learning cycles (less than 300 cycles in our experiments). The results also show that only a small portion of the state space was visited during the learning process, which supports the sparsity of the policy approximator grid (PAG). The main contributions of this work are as follows: 1) Development of an autonomous and unsupervised learning method that is designed for and effectively works for continuous state/action space control problems. 2) Development of data structures, which we call the policy approximator grid (PAG), the transition memory set (TMSET), and the action memory set (AMSET), that work seamlessly together with the proposed learning method. The insertion and retrieval times of these data structures are constant and independent of the amount of data that the data structures contain. 3) Experimental results that show that only a small portion of the state space is actually visited during the learning, as the state dimension increases. This "sparsity" allows the learning agent to govern the learning process without an exponentially increasing memory usage for the policy approximator implementation. 4) Development of a gradient descent trajectory optimization algorithm for phase-II learning. This algorithm is a general form of the 1st-order differential dynamic programming (DDP) algorithm.
Book
1 online resource.
Protein sorting coordinates membrane dynamics with the clustering and isolation of proteins into discrete membrane domains. Proper protein sorting not only mediates intracellular homeostasis, for example, through the delivery of hydrolase enzymes to the lysosome or through degradation of long-lived proteins in autophagy. It also facilitates the communication between a cell and its environment (the extracellular milieu or neighboring cells) by regulating both ligand secretion as well as the availability of receptors at the plasma membrane by balancing receptor degradation with receptor recycling. Disruption of protein sorting pathways can therefore interfere with intracellular processes and prevent the interactions between cells necessary for the maintenance of tissue homeostasis. The nervous system, in particular, is sensitive to disruption of protein sorting pathways, as evidenced by the growing number of reports linking polymorphisms in protein sorting machinery to neurological disease, such as Alzheimer's Disease (AD). It is therefore necessary to elucidate the mechanisms controlling protein sorting so that we may better understand the normal biology underlying these processes and the consequences of their disruption in pathogenesis. Previous work has established that levels of beclin 1, a protein that functions in multiple membrane trafficking pathways, are decreased in the brains of Alzheimer's Disease patients. We show that, in addition to it's well-known function in regulating autophagy, beclin 1 regulates receptor recycling in two systems relevant to neurodegeneration: phagocytosis in microglia and TGF-[beta] signaling in neurons. We use immunocytochemistry, confocal microscopy, and live-cell imaging to demonstrate beclin 1 regulates receptor recycling through production of phosphatidylinositol-3-phosphate at vesicle membranes and recruitment of the retromer complex. We also use a combination of biochemical techniques, flow cytometry, and immunohistochemistry to show the functional consequence of disrupted receptor recycling in phagocytosis and the TGF-[beta] signaling pathway using both in vitro and in vivo models. In light of our findings, we discuss the implications of impaired beclin 1-mediated protein sorting in neurological disease.
Protein sorting coordinates membrane dynamics with the clustering and isolation of proteins into discrete membrane domains. Proper protein sorting not only mediates intracellular homeostasis, for example, through the delivery of hydrolase enzymes to the lysosome or through degradation of long-lived proteins in autophagy. It also facilitates the communication between a cell and its environment (the extracellular milieu or neighboring cells) by regulating both ligand secretion as well as the availability of receptors at the plasma membrane by balancing receptor degradation with receptor recycling. Disruption of protein sorting pathways can therefore interfere with intracellular processes and prevent the interactions between cells necessary for the maintenance of tissue homeostasis. The nervous system, in particular, is sensitive to disruption of protein sorting pathways, as evidenced by the growing number of reports linking polymorphisms in protein sorting machinery to neurological disease, such as Alzheimer's Disease (AD). It is therefore necessary to elucidate the mechanisms controlling protein sorting so that we may better understand the normal biology underlying these processes and the consequences of their disruption in pathogenesis. Previous work has established that levels of beclin 1, a protein that functions in multiple membrane trafficking pathways, are decreased in the brains of Alzheimer's Disease patients. We show that, in addition to it's well-known function in regulating autophagy, beclin 1 regulates receptor recycling in two systems relevant to neurodegeneration: phagocytosis in microglia and TGF-[beta] signaling in neurons. We use immunocytochemistry, confocal microscopy, and live-cell imaging to demonstrate beclin 1 regulates receptor recycling through production of phosphatidylinositol-3-phosphate at vesicle membranes and recruitment of the retromer complex. We also use a combination of biochemical techniques, flow cytometry, and immunohistochemistry to show the functional consequence of disrupted receptor recycling in phagocytosis and the TGF-[beta] signaling pathway using both in vitro and in vivo models. In light of our findings, we discuss the implications of impaired beclin 1-mediated protein sorting in neurological disease.
Book
xvii, 678 pages : illustrations ; 24 cm.
Stanford University Libraries
Status of items at Stanford University Libraries
Stanford University Libraries Status
On order
(no call number) Unavailable On order Request
Book
1 online resource.
The key concept of Viktor Shklovsky's (1893-1984) understanding of literature is estrangement, a literary device the twofold purpose of which is to make a text strange and to restore one's perception of life. While previous scholarship has focused on the first of these purposes, this dissertation focuses on the second. Shklovsky, far from arguing for a vision of literature in which the meaning of a text is either nonexistent or located strictly in its language or formal features, in fact puts forth a theory of creativity as an existential enterprise, intimately connected with and even inseparable from deep human experiences such as alienation, love, transcendence, and the search for meaning. These experiences may be invoked or self-consciously created by language employed in specific, formally complex ways, but they can never finally be reduced to language. Particular attention is given to the way in which Shklovsky develops these ideas by engaging with the work of Roman Jakobson (1896-1982), Lev Tolstoy (1828-1910), and Vladimir Mayakovsky (1893-1930).
The key concept of Viktor Shklovsky's (1893-1984) understanding of literature is estrangement, a literary device the twofold purpose of which is to make a text strange and to restore one's perception of life. While previous scholarship has focused on the first of these purposes, this dissertation focuses on the second. Shklovsky, far from arguing for a vision of literature in which the meaning of a text is either nonexistent or located strictly in its language or formal features, in fact puts forth a theory of creativity as an existential enterprise, intimately connected with and even inseparable from deep human experiences such as alienation, love, transcendence, and the search for meaning. These experiences may be invoked or self-consciously created by language employed in specific, formally complex ways, but they can never finally be reduced to language. Particular attention is given to the way in which Shklovsky develops these ideas by engaging with the work of Roman Jakobson (1896-1982), Lev Tolstoy (1828-1910), and Vladimir Mayakovsky (1893-1930).
Book
vi, 260 pages : illustrations ; 23 cm.
Green Library
Status of items at Green Library
Green Library Status
Stacks Find it
(no call number) Unavailable On order Request
Book
448 pages ; 24 cm.
Stanford University Libraries
Status of items at Stanford University Libraries
Stanford University Libraries Status
On order
(no call number) Unavailable On order Request