In this thesis, a new smoothing penalty algorithm is introduced to solve a mathematical program with equilibrium constraints (MPEC). By smoothing the exact penalty function, an MPEC is reformulated as a series of subprograms which belong to a class of MPECs with simple linear complementarity constraints. To deal with the subproblems, a hybrid algorithm is proposed, which combines the active set algorithm, the 6-active search algorithm and the PSQP algorithm. It is shown that the smoothing penalty algorithm converges globally to a M-stationary point of MPEC under weak conditions. Supervisor: Dr. Jane Ye (Department of Mathematics and Statistics) Co-Supervisor: Dr. Wu-Sheng Lu (Department of Electrical and Computer Engineering)

Thesis (Ph. D.)--University of Washington, 1998. / Vita. Includes bibliographical references (leaves [104]-115).

Thesis (Ph. D.)--University of Washington, 2004. / Vita. Includes bibliographical references (p. 138-151).

Thesis (M.S.)--University of Texas at El Paso, 2007. / Title from title screen. Vita. CD-ROM. Includes bibliographical references. Also available online.

Call number: LD2668 .T4 1985 B86 / Master of Science

No description available.

In the first part of the thesis, we propose a new class of designs called multi-layer sliced Latin hypercube design (DSLHD) for running computer experiments. A general recursive strategy for constructing MLSLHD has been developed. Ordinary Latin hypercube designs and sliced Latin hypercube designs are special cases of MLSLHD with zero and one layer respectively. A special case of MLSLHD with two layers, doubly sliced Latin hypercube design, is studied in detail. The doubly sliced structure of DSLHD allows more flexible batch size than SLHD for collective evaluation of different computer models or batch sequential evaluation of a single computer model. Both finite-sample and asymptotical sampling properties of DSLHD are examined. Numerical experiments are provided to show the advantage of DSLHD over SLHD for both sequential evaluating a single computer model and collective evaluation of different computer models. Other applications of DSLHD include design for Gaussian process modeling with quantitative and qualitative factors, cross-validation, etc. Moreover, we also show the sliced structure, possibly combining with other criteria such as distance-based criteria, can be utilized to sequentially sample from a large spatial data set when we cannot include all the data points for modeling. A data center example is presented to illustrate the idea. The enhanced stochastic evolutionary algorithm is deployed to search for optimal design. In the second part of the thesis, we propose a new smoothing technique called completely-data-driven smoothing, intended for smoothing over irregular regions. The idea is to replace the penalty term in the smoothing splines by its estimate based on local least squares technique. A close form solution for our approach is derived. The implementation is very easy and computationally efficient. With some regularity assumptions on the input region and analytical assumptions on the true function, it can be shown that our estimator achieves the optimal convergence rate in general nonparametric regression. The algorithmic parameter that governs the trade-off between the fidelity to the data and the smoothness of the estimated function is chosen by generalized cross validation (GCV). The asymptotic optimality of GCV for choosing the algorithm parameter in our estimator is proved. Numerical experiments show that our method works well for both regular and irregular region smoothing. The third part of the thesis deals with uncertainty quantification in building energy assessment. In current practice, building simulation is routinely performed with best guesses of input parameters whose true value cannot be known exactly. These guesses affect the accuracy and reliability of the outcomes. There is an increasing need to perform uncertain analysis of those input parameters that are known to have a significant impact on the final outcome. In this part of the thesis, we focus on uncertainty quantification of two microclimate parameters: the local wind speed and the wind pressure coefficient. The idea is to compare the outcome of the standard model with that of a higher fidelity model. Statistical analysis is then conducted to build a connection between these two. The explicit form of statistical models can facilitate the improvement of the corresponding modules in the standard model.

This thesis consists of four chapters. Chapter 1 focuses on theoretical results on high-order laplacian-based regularization in function estimation. We studied the iterated laplacian regularization in the context of supervised learning in order to achieve both nice theoretical properties (like thin-plate splines) and good performance over complex region (like soap film smoother). In Chapter 2, we propose an innovative static path-planning algorithm called m-A* within an environment full of obstacles. Theoretically we show that m-A* reduces the number of vertex. In the simulation study, our approach outperforms A* armed with standard L1 heuristic and stronger ones such as True-Distance heuristics (TDH), yielding faster query time, adequate usage of memory and reasonable preprocessing time. Chapter 3 proposes m-LPA* algorithm which extends the m-A* algorithm in the context of dynamic path-planning and achieves better performance compared to the benchmark: lifelong planning A* (LPA*) in terms of robustness and worst-case computational complexity. Employing the same beamlet graphical structure as m-A*, m-LPA* encodes the information of the environment in a hierarchical, multiscale fashion, and therefore it produces a more robust dynamic path-planning algorithm. Chapter 4 focuses on an approach for the prediction of spot electricity spikes via a combination of boosting and wavelet analysis. Extensive numerical experiments show that our approach improved the prediction accuracy compared to those results of support vector machine, thanks to the fact that the gradient boosting trees method inherits the good properties of decision trees such as robustness to the irrelevant covariates, fast computational capability and good interpretation.

This work is dedicated to the control of the accuracy of computational simulations of sound propagation and scattering. Assuming time-harmonic behaviour, the mathematical models are given as boundary value problems for the Helmholtz equation <i>Delta u+k2u=0 </i> in <i>Oméga</i>. A distinction is made between interior, exterior and coupled problems and this work focuses mainly on interior uncoupled problems for which the Helmholtz equation becomes singular at eigenfrequencies. <p><p>As in other application fields, error control is an important issue in acoustic computations. It is clear that the numerical parameters (mesh size h and degree of approximation p) must be adapted to the physical parameter k. The well known ‘rule of the thumb’ for the h version with linear elements is to resolve the wavelength <i>lambda=2 pi k-1</i> by six elements characterising the approximability of the finite element mesh. If the numerical model is stable, the quality of the numerical solution is entirely controlled by the approximability of the finite element mesh. The situation is quite different in the presence of singularities. In that case, <i>stability</i> (or the lack thereof) is equally (sometimes more) important. In our application, the solutions are ‘rough’, i.e. highly oscillatory if the wavenumber is large. This is a singularity inherent to the differential operator rather than to the domain or the boundary conditions. This effect is called the <i>k-singularity</i>. Similarly, the discrete operator (“stiffness” matrix) becomes singular at eigenvalues of the discretised interior problem (or nearly singular at damped eigenvalues in solid-fluid interaction). This type of singularities is called the <i>lambda-singularities</i>. Both singularities are of global character. Without adaptive correction, their destabilizing effect generally leads to large error of the finite element results, even if the finite element mesh satisfies the ‘rule of the thumb’. <p><p>The k- and lambda-singularities are first extensively demonstrated by numerical examples. Then, two <i>a posteriori</i> error estimators are developed and the numerical tests show that, due to these specific phenomena of dynamo-acoustic computations, <i>error control cannot, in general, be accomplished by just ‘transplanting’ methods that worked well in static computations</i>. However, for low wavenumbers, it is necessary to also control the influence of the geometric (reentrants corners) or physical (discontinuities of the boundary conditions) singularities. An <i>h</i>-adaptive version with refinements has been implemented. These tools have been applied to two industrial examples :the GLT, a bi-mode bus from Bombardier Eurorail, and the Vertigo, a sport car from Gillet Automobiles.<p><p>As a conclusion, it is recommanded to replace the rule of the thumb by a criterion based on the control of the influence of the specific singularities of the Helmholtz operator. As this aim cannot be achieved by the <i>a posteriori</i> error estimators, it is suggested to minimize the influence of the singularities by modifying the formulation of the finite element method or by formulating a “meshless” method.<p> / Doctorat en sciences appliquées / info:eu-repo/semantics/nonPublished