Search

China has been experiencing rapid economic growth over the past few decades. Accompanied by the nation's modernization and urbanization, an overwhelmingly large number of rural habitants out-migrate to urban areas for work and better earnings. However, due to China's local household registration system(Hukou), children of migrants are not entitled to free schooling provisions in these urban destinations. As a consequence, millions of school-agedrural children have to be 'left behind' by their migrating parent(s) in their homeplace. They are called - the 'left-behind children'. This study investigates the future educational aspirations of Chinese 'left-behind children' and explores the educational impact of 'left-behind-ness'. It adoptsthe qualitative case study approach and uses ethnography as the primary research technique to capture factors that impact on aspirations, and to understand and explain why educational aspirations are different among these young people who share many similar social characteristics. Fieldwork was conducted in a nine-year comprehensive school in a rural region in southwest China, with data mainly being obtained from 17 'left-behind children'. Findings indicate that whilst undeniably educational aspirations are embedded in left-behind children's distinctly disadvantaged social background, aspirations are determined by these young people's own evaluation of whether attending university is 'rational', and are shaped by their individual family and school experience, relationship with adults and peers, perceived sense of self, as well as their personal and varied understandings of the social world they are living in. By focusing on the individuals and their experiences, this study suggests that being 'left behind' is experienced individually and it is these individual experiences that have influenced attitudes and aspirations differently among Chinese 'left-behind children'. This study not only extends understanding of Chinese leftbehind children's educational experiences, but also provides new knowledge regarding Chinese left-behind children's educational aspirations with a contribution to the integration of theoryas well as the broadening of theoretical applications by transferring Western theories to an Eastern context.

This research offers novel insights into how supply chain partnerships can be managed through voluntarily adopted non-proprietary based supply chain management (SCM) practices. This research is also timely as reflected by the emergence of these types of practices in recent times, leading to the suggestion that they might be equally effective as (if not more so than) traditional proprietary based ones in managing supply chain partnerships. Previous research has tended to focus on proprietary based SCM practices that are meant to be mandatorily adopted among supply chain firms. It has however tended to focus on the transaction itself, detracting attention from the broader issue of understanding how partnerships are managed as such. The lack of consideration of the relational aspects of partnerships is especially made prominent given that fums these days rarely function as autonomous entities but are in some form of partnership with others. The lack of any established theory therefore suggests a theory-building as opposed to a theory-testing research approach. Through case studies analysing the behaviour of firms in supply chain partnerships upon the introduction of voluntarily adopted non-proprietary SCM practices, insights were revealed into how partnerships were managed as such. This research found that these practices also provided an opportunity for a multitude of behaviours to be exhibited, which might not otherwise be possible in supply chain partnerships managed by mandatorily adopted proprietary based SCM practices. Specifically, it found that firms responded by undertaking various governance and adaptation activities. It also found that these two activities were related to each other. In the broader scheme of things, it is suggested that through these types of practices, firms are able to meet the challenge of managing their supply chain partnerships. The key implications for theory and strategy are: (1) it challenges the predominant view that voluntarily adopted non-proprietary based SCM practices are ill-equipped to manage supply chain partnerships and, (2) the recognition that the behaviour of firms in supply chain partnerships is affected by interdependency effects.

In this paper, we tackle machine learning over molecular space by considering three representations for molecules: (1) a vector of molecular properties that we treat as predictor variables, (2) a graph that captures the relationship between individual atoms in a molecule, and (3) a cheminformatic fingerprint that “identifies” a molecule. We assess the viability of each representation by training a model to predict energy values. In particular, we look a class of models that use kernel methods, whereby the prediction algorithm relies on a similarity measure between training data. On a subset of the Harvard Clean Energy Project (CEP) database, we find a simple fingerprint similarity kernel to be the fastest and most accurate for predicting HOMO-LUMO energy gap values.

This thesis outlines a formalism to study various physical phenomena based on a generalized concept of quaternions. It is shown that there is a total of 64 possible multiplication rules which can be defined starting with the generalized imaginary units first introduced by Hamilton. Of these 64 choices, only 8 lead to non-commutative division algebras, two of which are associated to the left- and right-chirality quaternions while the other six are generalizations of the split-quaternion, a concept first introduced by Cockle. Next, we show that both the left- and right- chirality versions of the generalized split-quaternions in their 4 × 4 matrix forms are algebraically isomorphic and can be related to each other by 4 × 4 permutation matrices of the C2 × C2 group. We first show that the left- and right-chirality quaternions can be used (1) to generate clockwise and anti-clockwise rotations in our three-dimensional space, (2) to describe Lorentz boosts with constant velocity in an arbitrary direction, (3) to introduce a new approach to calculate the probability of spin components, as well as the mean and standard deviation of a Hermitian operator in quantum mechanics, and (4) to reformulate quantum gates and qubits in quantum computing. We then use the quaternion formalism to study noise in quantum circuits and calculate the qubit fidelity after applications of many noisy gates on an initial qubit state. Next, it is shown how each of the generalized split-quaternion algebras can be used to solve the problem of quantum-mechanical tunneling through an arbitrary one-dimensional conduction band energy profile. This demonstrates that six different spinors (4 × 4 matrices) can be used to represent the amplitudes of the left and right propagating waves in a device. Finally, we point out an interesting connection between the non-commutative split-quaternion division algebras and the uniqueness of our four-dimensional space-time by introducing a super split-quaternion formed of a superposition of either the left- or right-chirality versions of split-quaternions. It is shown that the norm of this super split-quaternion has a form identical to the space-time interval. This opens the door for an interpretation of the special and general theories of relativity based on a quaternionic viewpoint of the universe.

During the last 30 years, microelectronic devices have been continuously designed and developed with smaller size and yet more functionalities. Today, hundreds of millions of transistors and complementary metal-oxide-semiconductor cells can be designed and integrated on a single microchip through 3D packaging and chip stacking technology. A large amount of heat will be generated in a limited space during the operation of microchips. Moreover, there is a high possibility of hot spots due to non-uniform integrated circuit design patterns as some core parts of a microchip work harder than other memory parts. This issue becomes acute as stacked microchips get thinner. In other applications, laser devices can generate heat fluxes up to 1000 W/cm2 in less than 0.5 mm2 areas. Light-emitting diodes also entail high heat intensities between 300 and 600 W/cm2 due to extremely high power density. Therefore, it is of technological importance that heat dissipation can be well managed and controlled in microelectronics devices. This thesis is mainly focused on the micro/nanoscale thermal conductivity and interfacial thermal resistance characterization and optimization in two-dimensional (2D) nanostructures, such as graphene, C2N, C3N, phosphorene, stanene, molybdenum disulfide, and molybdenum diselenide. Various approaches including non-equilibrium molecular dynamics (NEMD) simulation, equilibrium molecular dynamics (EMD) simulation, and transient pump-probe approaches have been utilized to explore the thermal properties. Phonon behaviors have also been studied to explain the mechanism of heat transfer. Then various machine learning (ML) models such as linear regression, polynomial regression, decision tree, random forest, and artificial neural network have been employed to predict the thermal properties of 2D materials. In a different area of research, the water desalination performance of carbon nanotube with rim functionalization has been systematically investigated using molecular dynamics (MD) simulations. Advisor: Xiao Cheng Zeng

Normal (uncut) tensile strength is weakly dependent on the concentration of carbon black N330. Strain at break has decreases as the concentration of N330 increases. 100% modulus increases as the concentration of N330 increases.With less than 15 phr of N330, NR vulcanizates were weaker than the gum, and tensile strength decreased as the amount of N330 increased. All gum specimens, NC3 and NC6 show single lateral cracking.With 15 phr-17 phr of N330, a fraction of specimens had larger strength than the gum, and multiple cracks occurred before rupture.With 18 phr-40 phr of N330, NR vulcanizates were stronger than the gum. Tensile strength increased as the amount of N330 increased. All specimens show multiple cracks prior to rupture.N330 sulfur-cured vulcanizates form a coherent gel when the carbon black concentration is higher than 16 phr.The overall reinforcement is in the range of 17phr-18phr. Compared to the similar vulcanizates filled with carbon black N660, N330-filled vulcanizates have a lower onset concentration for reinforcement, bound rubber formation, and they have larger 100% modulus and tensile strength for both uncut and edge-cut specimens.

Uninhabited Aerial Vehicle (UAV) is becoming increasingly attractive in missions where human presence is undesirable or impossible. Agile maneuvers and long endurance are among the most desired advantages of UAVs over aircraft that have human pilots onboard. Past studies suggest that the performance of UAVs may be considerably improved by utilizing natural resources, especially wind energy, during flights. The key challenge of exploiting wind energy in practical UAV operations lies in the availability of reliable and timely wind field information in the operational region. This thesis presents a practical onboard strategy that attempts to overcome this challenge, to enable UAVs in utilizing wind energy effectively during flights, and therefore to enhance performance. We propose and explore a strategy that combines wind measurement and optimal trajectory planning onboard UAVs. During a cycle of a loiter flight, a UAV can take measurements of wind velocity components over the flight region, use these measurements to estimate the local wind field through a model-based approach, and then compute a flight trajectory for the next flight cycle with the objective of optimizing fuel. As the UAV follows the planned trajectory, it continues to measure the wind components and repeats the process of updating the wind model with new estimations and planning optimal trajectories for the next flight cycle. Besides presenting an onboard trajectory planning strategy of wind energy exploration, estimation, and utilization, this research also develops a semi-analytical linearized solution to the formulated nonlinear optimal control problem. Simulations and numerical results indicate that the fuel savings of trajectories generated using the proposed scheme depend on wind speed, wind estimation errors, rates of change in wind speed, and the wind model structures. For a given wind field, the magnitude of potential fuel savings is also contingent upon UAVs' performance capabilities.