Microstructure or fabric definitely affects macroscopic mechanical behavior of granular material. It is also well-observed that fabric evolves with shearing or plastic deformation. In this study, a series of two-dimensional numerical direct shear tests are carried out with the discrete element method, to study the initial fabric effect on global material responses and their micro-macroscopic relations. Idealized particle assemblages are made up of mono-size elongated particles and are prepared by a “deposition” method. Elongated particle is modeled by the built-in clump logic, in which constitutive balls are joined together without further breakage. In the deposition method, there are three controlling parameters, including, deposited direction, inter-particle friction coefficient and particle number, to prepare specimens with similar initial density but different initial packing or fabric. Three types of fabric of particle assemblages are examined quantitatively and are monitored during shearing, including, particle orientations (PO), contact normal forces (NF), and void spaces (VS). These fabric distributions are described by two parameters―anisotropic degree ( ) and orientation angle ( ), with clear physical implications. An additional parameter ( ) describing the average size of voids, is used to quantify void perimeter. It is found that this parameter has a relation with the assemblage’s volumetric response. C With the systematic and meticulous quantification method, the linkage between the macroscopic and microscopic responses of particle assemblages is discussed quantitatively. The results show that the initial packing affects the shear zone thickness, initial stiffness, peak strength, and dilation rate. In the shear zone, particle orientations do not exhibit a unique state at the final stage of direct shearing. At that state, strong normal forces and strong voids are parallel to the major principal stress direction. It seems that the initial packing does not affect their final distributions. At the end of reverse shearing, strong voids and strong normal forces in the shear zone give an essentially unique state, and their preferential directions are related to the changed loading direction. However, apparent stable particle orientations are still affected by the initial fabric. / published_or_final_version / Civil Engineering / Master / Master of Philosophy

Logistics makes extensive use of human and material resources to achieve a target level of customer service at the lowest possible cost. It has been recognized as a major key to success in commerce and industry, and continues to evolve radically and grow in importance in recent years. Warehousing, as one of the most costly elements of logistics, is often the central operation in most logistics networks. Its successful management is critical in terms of both cost and service. In this thesis, two problem areas in warehouse logistics are studied: stock loading and order picking. Stock loading is an essential operation in modern logistics. Improvement on container capacity utilization and loading efficiency significantly reduces costs. For a given set of boxes in different sizes and an unlimited number of identical containers, the basic cargo loading problem is to determine the minimum number of containers required. The problem is proven NP-hard. To tackle this problem, a Tabu search optimization with a tree-based cargo loading algorithm as its inner heuristic is proposed. This approach has flexibility in taking different box conditions into consideration, and can find better solutions on average than other recent meta- or heuristic algorithms. Decreasing order sizes and increasing fuel costs provide a strong incentive for the inner-city truck loading operation to utilize container space more efficiently in transporting goods to multiple clients during one trip. This considers not only traditional loading constraints, but also multi-drop requirements. A wallbuilding heuristics based on a binary tree data structure is proposed to handle these side constraints. A dynamic space decomposition approach, together with a repacking and space amalgamation strategy, permits an efficient and effective loading plan. Order picking, one of the most critical warehousing operations, is the second problem studied in this thesis. An analytical approximation model is proposed based on probability modeling and queueing network theory applied to a synchronized zone picker-to-part order picking system with different routing and ABC-class inventory storage policies. The numerical results are compared and validated via simulation. The resulting model can therefore be usefully applied in the design and selection process of order picking systems. The routing versus storage issues are further investigated with a simulation model. This extends the existing research by evaluating multiple routing and storage policies under varying operating conditions. Results show that the midpoint, return and traversal routing policies generally perform best when paired with perimeter, across-aisle and within-aisle storage strategies, respectively. Yet performance is indeed dependent on demand patterns, zone sizes, batch sizes and order sizes. At first glance, order picking and stock loading operation seem to pursue different objectives. However, they are two related operations conducted sequentially from internal to the outbound side of warehousing. An efficient order picking system is a precondition for an effective loading operation at the shipping dock, especially when multiple orders need to be selected for consolidation in shipment. The proposed loading algorithms and the order picking system performance evaluation models can be used to further study the effective integration of these two functions. / published_or_final_version / Mathematics / Doctoral / Doctor of Philosophy

Supplier selection and order allocation are significant decisions for a manufacturer to ensure stable material flows in a highly competitive supply chain, in particular when customers are willing to accept products with less desirable product attributes. Hence, this study develops efficient methodologies to solve optimally the integrated supplier selection and order allocation problem incorporating customer flexibility for a manufacturer producing multiple products over a multi-period planning horizon. In this research, a new fuzzy multi-attribute approach is proposed to evaluate customer flexibility which is characterized through range and response. The approach calculates the product’s general utility value. This value is used by a bi-variant function which is developed to determine the retail price for the product. A new mixed integer program model describing the behavior of the basic problem is firstly developed. This basic model is the first to jointly determine: 1) type and quantity of the product variants to be offered; 2) the suppliers to be selected and orders to be allocated; and 3) inventory levels of product variants and raw materials/components. The objective is to maximize the manufacturer’s total profit subject to various operating constraints. This basic problem constitutes a very complex combinatorial optimization problem that is Nondeterministic Polynomial (NP)-hard. To tackle this challenge, two new optimization algorithms, i.e., an improved genetic approach called king GA (KGA) and an innovative hybrid algorithm called (CP-SA) _I which combines the techniques of constraint programming and simulated annealing are developed to locate optimal solutions. Extensive computational experiments demonstrate the effectiveness of these algorithms and also show clearly that (CP-SA) _I outperforms KGA in terms of both solution quality and computational cost. To examine the influence of subcontracting as one widespread practice in modern production management, this study also develops a modified mathematical model. It shares some similarity with the basic model but brings additional complexity by taking into consideration subcontractors for inter-mediate components and machine capacity. Since (CP-SA) _I outperforms KGA, it is employed and modified to solve the modified problem. Hence, this study presents a new hybrid algorithm called (CP-SA) _II, to locate optimal solutions. This study also establishes a new parallel (CP-SA) _II algorithm to enhance the performance of (CP-SA) _II. This parallel algorithm is implemented on a distributed computing platform based on the contemporary Graphic Processing Unit (GPU) using the Compute Unified Device Architecture (CUDA) programming model. Extensive numerical experiments conducted clearly demonstrate that the parallel (CP-SA) _II algorithm and its serial counterpart are efficient and robust optimization tools for formulating integrated supplier selection and order allocation decisions. Sensitivity analysis is employed to study the effects of the critical parameters on the performance of these algorithms. Finally, the convergence behavior of the proposed parallel (CP-SA) _II algorithm is studied theoretically. The results prove that the search process eventually converges to the global optimum if the overall best solution is maintained over time. / published_or_final_version / Industrial and Manufacturing Systems Engineering / Doctoral / Doctor of Philosophy

Research conducted in mathematical finance focuses on the quantitative modeling of financial markets. It allows one to solve financial problems by using mathematical methods and provides understanding and prediction of the complicated financial behaviors. In this thesis, efforts are devoted to derive and extend stochastic optimization models in financial economics and establish practical algorithms for representing and solving problems in mathematical finance. An option gives the holder the right, but not the obligation, to buy or sell an underlying asset at a specified strike price on or before a specified date. In this thesis, a valuation model for a perpetual convertible bond is developed when the price dynamics of the underlying share are governed by Markovian regime-switching models. By making use of the relationship between the convertible bond and an American option, the valuation of a perpetual convertible bond can be transformed into an optimal stopping problem. A novel approach is also proposed to discuss an optimal inventory level of a retail product from a real option perspective in this thesis. The expected present value of the net profit from selling the product which is the objective function of the optimal inventory problem can be given by the actuarial value of a real option. Hence, option pricing techniques are adopted to solve the optimal inventory problem in this thesis. The goal of risk management is to eliminate or minimize the level of risk associated with a business operation. In the risk measurement literature, there is relatively little amount of work focusing on the risk measurement and management of interest rate instruments. This thesis concerns about building a risk measurement framework based on some modern risk measures, such as Value-at-Risk (VaR) and Expected Shortfall (ES), for describing and quantifying the risk of interest rate sensitive instruments. From the lessons of the recent financial turmoils, it is understood that maximizing profits is not the only objective that needs to be taken into account. The consideration for risk control is of primal importance. Hence, an optimal submission problem of bid and ask quotes in the presence of risk constraints is studied in this thesis. The optimal submission problem of bid and ask quotes is formulated as a stochastic optimal control problem. Portfolio management is a professional management of various securities and assets in order to match investment objectives and balance risk against performance. Different choices of time series models for asset price may lead to different portfolio management strategies. In this thesis, a discrete-time dynamic programming approach which is flexible enough to deal with the optimal asset allocation problem under a general stochastic dynamical system is explored. It’s also interesting to analyze the implications of the heteroscedastic effect described by a continuous-time stochastic volatility model for evaluating risk of a cash management problem. In this thesis, a continuous-time dynamic programming approach is employed to investigate the cash management problem under stochastic volatility model and constant volatility model respectively. / published_or_final_version / Mathematics / Doctoral / Doctor of Philosophy

Numerical modeling using finite element method (FEM) is well-recognized as a powerful method for both engineers and researchers to solve boundary value problems. In the modeling of geotechnical problems, the analyses are often limited to simple static problems with either steady-state effective or total stress approach while the transient response (development and dissipation of excess pore water pressure, uex) is seldom considered. Besides, infinitesimal small soil deformation is usually assumed. The simulation is further complicated when the soil-structure interaction problems involve significant soil displacements; like a pile subject to negative skin friction (NSF) and a cone/pile penetration. However, conventional FEM analysis prematurely terminates due primarily to excessive mesh distortion. One could see that simulating a transient problem with large deformation and distortion remains a great challenge. In this study, advanced FE simulations are performed to give new insights into the problems of (1) a pile subject to NSF; and (2) a cone penetration. The transient response of the NSF problem is modeled with the fluid-coupled consolidation technique and geometric nonlinearity. The fluid-coupled cone penetration problem is modeled with a newly developed adaptive approach. The NSF and cone penetration simulations involve complex soil-structure interface modeling. Two types of modified interface responses are developed and verified which consider fluid coupling. The developed algorithm is applied to back analyze a case history of a pile subject to NSF induced by surcharge loading. Promising results were shown. Development of dragload and neutral plane (NP) with time is studied. NP locates at 75% of the pile embedded length (D) in long-term. Next, a parametric study is performed to investigate the influences of pile geometries, ground compressibility and loading conditions towards the pile responses. The long-term NP locates at around 0.55D to 0.65D in the studied engineering scenarios. The maximum downdrag can be up to 10% of the pile diameter. NP shifts upward when the head load increases. A simple design chart is proposed which helps engineers to estimate the long-term axial load distribution. An illustrative example is given to demonstrate the application and performance of the chart. The study is extended to investigate the cone penetration problem. An advanced adaptive method is developed and implemented into the FE package ABAQUS to resolve the problems of numerical instability, excessive mesh distortion and premature termination. The proposed method is verified by modeling a ground consolidation problem. Next, total stress back analysis of cone penetration is conducted with the proposed method. The development of cone factor predicted by the proposed method gives a better match with the laboratory result when comparing with the built-in ALE method. Next, the development and dissipation of uex during cone advancing with the proposed method and fluid-coupled technique is investigated. uex develops dramatically around the cone tip. The soil permeability is back calculated from the dissipation test and agrees well with the input value. It is believed that the construction effects of a press-in pile and the subsequence NSF on that pile can be modeled by utilizing the finding of this study. / published_or_final_version / Civil Engineering / Doctoral / Doctor of Philosophy

Background: The seriousness of human influenza virus infection, in combination with the transmissibility of the virus, determines the impact that the virus will have in a population. However, the uncertainty surrounding the seriousness and changes in seriousness hindered the calibration of the early public health response. Methods: I applied statistical models to population-based mortality data and hospitalizations among patients with laboratory-confirmed H1N1pdm09. I estimated the disease burden, retrospectively and prospectively determined seriousness of influenza virus infections including the risk of death on a per-infection basis (IFR) and on a per-hospitalization (HFR) of H1N1pdm09, and investigated changes in seriousness. I used serologic surveillance data to estimate the cumulative incidence of infection in a population, and used it as the denominator of the IFR. I also conducted systematic reviews and meta-analysis to summarize published estimates of the risk of death among cases (CFR) and HFR of the pandemic influenza H1N1pdm09 virus. Results: I estimated that the first wave of H1N1pdm09 was associated with approximately 232 (95% confidence interval: 136, 328) excess deaths in all ages in Hong Kong, which was around 4 times the observed number of laboratory-confirmed deaths of H1N1pdm09. The point estimates for the IFR and HFR increased substantially with age. I included 77 estimates for the CFR from 50 published studies; whereas I included 187 estimates for the HFR from 184 published studies. The CFR was widely used to assess seriousness but the variation of a ‘case’ varied considerably in the literature. Variability in published estimates of the HFR was much less than variability in the CFR. Conclusions: Early in the next pandemic, estimation of the HFR, IFR or symptomatic CFR may provide a timely picture of the seriousness of infection, particularly if presented in comparison between two influenza virus infections in the same setting. Ongoing monitoring of mortality and influenza activity could permit identification of changes in seriousness of influenza virus infections. / published_or_final_version / Public Health / Doctoral / Doctor of Philosophy

Not available / text

Not available / text

Not available / text

Not available / text