In this thesis we focus on modelling correlation between selected stock markets using high-frequency data. We use time-series of returns of following indices: FTSE, DAX PX and S&P, and Gold and Oil commodity futures. In the first part of our empirical work we compute daily realized correlations between returns of subject instruments and discuss the dynamics of it. We then compute unconditional correlations based on average daily realized correlations and using feedforward neural network (FFNN) to assess how well the FFNN approximates realized correlations. We also forecast daily realized correlations of FTSE:DAX and S&P:Oil pairs using heterogeneous autoregressive model (HAR), autoregressive model of order p (AR(p)) and nonlinear autoregressive neural network (NARNET) and compare performance of these models.

This thesis analyses, derives and evaluates specification tests of Generalized Auto-Regressive Conditional Heteroskedasticity (GARCH) regression models, both univariate and multivariate. Of particular interest, in the first half of the thesis, is the derivation of robust test procedures designed to assess the Constant Conditional Correlation (CCC) assumption often employed in multivariate GARCH (MGARCH) models. New asymptotically valid conditional moment tests are proposed which are simple to construct, easily implementable following the full or partial Quasi Maximum Likelihood (QML) estimation and which are robust to non-normality. In doing so, a non-normality robust version of the Tse's (2000) LM test is provided. In addition, a new and easily programmable expressions of the expected Hessian matrix associated with the QMLE is obtained. The finite sample performances of these tests are investigated in an extensive Monte Carlo study, programmed in GAUSS.In the second half of the thesis, attention is devoted to nonparametric testing of GARCH regression models. First simultaneous consistent nonparametric tests of the conditional mean and conditional variance structure of univariate GARCH models are considered. The approach is developed from the Integrated Generalized Spectral (IGS) and Projected Integrated Conditional Moment (PICM) procedures proposed recently by Escanciano (2008 and 2009, respectively) for time series models. Extending Escanciano (2008), a new and simple wild bootstrap procedure is proposed to implement these tests. A Monte Carlo study compares the performance of these nonparametric tests and four parametric tests of nonlinearity and/or asymmetry under a wide range of alternatives. Although the proposed bootstrap scheme does not strictly satisfy the asymptotic requirements, the simulation results demonstrate its ability to control the size extremely well and therefore the power comparison seems justified. Furthermore, this suggests there may exist weaker conditions under which the tests are implementable. The simulation exercise also presents the new evidence of the effect of conditional mean misspecification on various parametric tests of conditional variance. The testing procedures are also illustrated with the help of the S&P 500 data. Finally the PICM and IGS approaches are extended to the MGARCH case. The procedure is illustrated with the help of a bivariate CCC-GARCH model, but can be generalized to other MGARCH specifications. Simulation exercise shows that these tests have satisfactory size and are robust to non-normality. The marginal mean and variance tests have excellent power; however the covariance marginal tests lack power for some alternatives.

Digital Laser Speckle Image Correlation (DiLSIC) is a technique that utilizes a laser generated speckle pattern with Digital Image Correlation (DIC). This technology eliminates the need to apply an artifact speckle pattern to the surface of the material of interest, and produces a finer speckle pattern resulting in a more sensitive analysis. This investigation explores the parameters effecting laser speckle patterns for DIC and studies DiLSIC as a tool to measure surface strain and detect subsurface defects on pressure vessels. In this study a 632.8 nm 30 mW neon-helium laser generated the speckle pattern by passing through the objective end of an objective lens. All experiments took place in a lab setting on a high performance laminar flow stabilizer optical table.This investigation began with a deeper look at the camera settings that effect the effectiveness of using laser speckles with DIC. The first studies were concentrated on the aperture size (f-stop), shutter speed, and gain (ISO) of the camera. Through a series of zero-correlation studies, translation tests, and settings studies, it was discovered that, much like white light DIC, an increased gain allowed for more noise and less reliable measurements when using DiLSIC. It was shown that the aperture size and shutter speed will largely depend on the surface composition of the material, and that these factors should be investigated with each new sample of different surface finish.To determine the feasibility of using DiLSIC on pressure vessels two samples were acquired. The first was a standard ASTM filament wound composite pressure vessel (CPV) which had an upper load limit of 40 psi. The second was a plastic vessel that had internal subsurface defects added with the use of an air pencil grinder. Both vessels were put under a pressure load with the use of a modified air compressor that allowed for multiple loading cycles through the use of a pressure relief valve. The CPV was mapped out in 10-degree increments between the 90° and 180° markings that were on the pressure vessel, occurring in three areas, each one inch apart. The CPV had a pressure load applied to at 10, 20, 30,and 40 psi. DiLSIC was able to measure increasing displacement with increased loading on the surface of the CPV, however with a load limit of 40 psi no strains were detected. The plastic vessel had known subsurface defects, and these areas were the focus of the investigation. The plastic vessel was loaded with a pressure load at 5, 10, 12, 15, 17, and 20 psi. The 5 psi loaded image was used as a reference image for the correlation and decorrelation consistently occurred at 20 psi. This investigation proved that DiLSIC can detect and locate subsurface defects through strain measurement. The results were verified with traditional white light DIC, which also showed that the subsurface defects on pressure vessels were detectable. The DIC and DiLSIC results did not agree on maximum strain measurement, with the DiLSIC prediciting much larger strains than traditional DIC. This is due to the larger effect out-of-plane displacement has on DiLSIC. DiLSIC was able to detect subsurface defects on a pressure vessel. The median measured hoop strain was in agreement for DiLSIC, DIC and the predicted hoop strain for a wall thickness of 0.1 inches. However, DiLSIC also produced unreliable maximum strain measurements. This technique shows potential for future applications, but more investigations will be needed to implement it for industrial use. A full investigation into the parameters surrounding this technique, and the factors that contribute the most to added noise and unreliability should be conducted. This technology is being developed by multiple entities and shows promising results, and once further advanced could be a useful tool for rapid surface strain measurement and subsurface defect detection in nondestructive evaluation applications. Therefore, it is recommended to continue further investigations into this technology and its applications.

<div>Due to reduced pain, shortened hospital stay and recovery, minimally invasive surgery (MIS) is becoming more and more popular in healthcare systems. MIS requires some devices for the motion and force transformation from outside to the inside of the body of a patient, and the strangled cables play a significant role in developing the instrumentations to serve for such purposes. However, current design and selection of a strangled cable is mostly intuitive that depend greatly on designers’ experiences and availability of experimental data, which leads to non-optimized designs and longer design cycles. In this thesis, both of analytical modelling and numerical simulation are proposed to build the relation of applied torque and deflection of part, so that a strangled cable with a given configuration can be characterized in term of its load-deflection relation. The defined relation has its great significance and application potential in the design optimization and precise controls of medical devices for MISs.Besides the various patterns of strangled cables, a Helical hollow strand (HHS®) tube is a special type of strangled cables with single- or multiple- layer configurations., In each layer, each of the helical wires touches its two neighboring helical wires, and it has a coreless hollow at the center. Its primary application is to carry a torsional load in a twisting mode. As an extreme, there is a possibility that all helical wires touch each other, and this forms a statically indeterminate contact obstacle in design analysis. Numerical simulation would predict that contacts occur simultaneously at all possible contacting points under the circumference that the strand is fixed at one end against rotation. In addition, the friction at contacts will affect the torsional deformation; therefore, these contacts must be taken into consideration in the development of analytical and numerical simulation models.This thesis reports the results of the investigation on the characteristics of Helical hollow strand tube (HHS®), more specifically, the relation of torsional deflection and the applied torque over a tube in the clockwise (CW) direction. The numerical simulation approach to predict the torsional deflection of HHS with various design parameters and configurations is emphasized. </div><div><br></div>

計算機模型可以描述複雜的物理現象，然而這些模型應用在科學研究時其運算需要很長的時間，而且要有特定的實驗設計才能了解現象的本質。在有缺少一個或數個主效應的情形下，因子設計是不適合的，因為在缺少主效應下時其重複實驗不但不能估計誤差，只是產生重複實驗。雖然已經有學者提出許多可替代的設計，但是大部份設計的計算還是很累贅。本篇論文所提出的一些設計是從旋轉平面的二維因子設計發展而來，這些旋轉因子的設計很容易建構而且保有許多標準因子設計中吸引人的性質：（1）在每個維度的投影是均等空間投影；（2）在迴歸模型中，估計效應是不相關的（即正交的）。這些設計被稱為最大化最小拉丁超方陣，其設計與近期學者建構的最小化內點間距離的準則是同等的。 關鍵字：有效相關（(Effect Correaltion）、拉丁超方陣（Latin Hypercube）、最大化最小距離（Maximin Distance）、最大化最小內點距離（Minimum Interpoint Distance） / Computer models can describe complicated physical phenomena. To use these models for scientific investigation, however, their generally long running times and mostly deterministic nature require a special designed experiment. Standard factorial designs are inadequate; in the absence of one or more main effects, their replication cannot be used to estimate error but instead produces redundancy. A number of alternative designs have been proposed, but many can be burdensome computationally. This paper presents a class of designs developed from the rotation of a two-dimensional factorial design in the plane. These rotated factorial designs are very easy to construct and preserve many of the attractive properties of standard factorial designs: they have equally-spaced projections to univariate dimensions and uncorrelated regression effect estimates (orthogonality) . They also rate comparably to maximin Latin hypercube designs by the minimum interpoint distance criterion used in the latter "s construction. Key Word : Effect Correlation, Latin Hypercube, Maximin Distance, Minimum Interpoint Distance