No description available.

In this dissertation we present new results regarding optimality of block designs with limited resources. The dissertation is organized as follows. The first chapter outlines the theory of optimal block design. The second chapter shows new work in optimal minimally connected block designs with spatial correlation structure. The third chapter details the discovery of the optimal incomplete designs with two blocks. The fourth chapter does the same for the optimal binary incomplete designs with three blocks. The fifth chapter summarizes the techniques used and new results found and lists possible future research topics. / Ph. D.

No description available.

No description available.

We consider optimal design for changepoint problems with particular attention paid to situations where the only possible change is in the mean. Optimal design for changepoint problems has only been addressed in an unpublished doctoral thesis, and in only one journal article, which was in a frequentist setting. The simplest situation we consider is that of a stochastic process that may undergo a, change at an unknown instant in some interval. The experimenter can take n measurements and is faced with one or more of the following optimal design problems: Where should these n observations be taken in order to best test for a change somewhere in the interval? Where should the observations be taken in order to best test for a change in a specified subinterval? Assuming that a change will take place, where should the observations be taken so that that one may best estimate the before-change mean as well as the after-change mean? We take a Bayesian approach, with a risk based on squared error loss, as a design criterion function for estimation, and a risk based on generalized 0-1 loss, for testing. We also use the Spezzaferri design criterion function for model discrimination, as an alternative criterion function for testing. By insisting that all observations are at least a minimum distance apart in order to ensure rough independence, we find the optimal design for all three problems. We ascertain the optimal designs by writing the design criterion functions as functions of the design measure, rather than of the designs themselves. We then use the geometric form of the design measure space and the concavity of the criterion function to find the optimal design measure. There is a straightforward correspondence between the set of design measures and the set of designs. Our approach is similar in spirit, although rather different in detail, from that introduced by Kiefer. In addition, we consider design for estimation of the changepoint itself, and optimal designs for the multipath changepoint problem. We demonstrate why the former problem most likely has a prior-dependent solution while the latter problems, in their most general settings, are complicated by the lack of concavity of the design criterion function. / Nous considérons, dans cette dissertation, les plans d'expérience bayésiens optimauxpour les problèmes de point de rupture avec changement d'espérance. Un cas de pointde rupture avec changement d'espérance à une seule trajectoire se présente lorsqu'uneséquence de données est prélevée le long d'un axe temporelle (ou son équivalent) etque leur espérance change de valeur. Ce changement, s'il survient, se produit à unendroit sur l'axe inconnu de l'expérimentateur. Cet endroit est appelé "point derupture". Le fait que la position du point de rupture soit inconnue rend les tests etl'inférence difficiles dans les situations de point de rupture à une seule trajectoire.

The optimal projection approach to solving the reduced order model problem produces two coupled, highly nonlinear matrix equations with rank conditions as constraints. Due to the resemblance of these equations to standard matrix Lyapunov equations, they are called modified Lyapunov equations. The proposed algorithms utilize probability-one homotopy theory as the main tool. It is shown that there is a family of systems (the homotopy) that make a continuous transformation from some initial system to the final system. With a carefully chosen initial problem a theorem guarantees that all the systems along the homotopy path will be asymptotically stable, controllable and observable. One method, which solves the equations in their original form, requires a decomposition of the projection matrix using the Drazin inverse of a matrix. It is shown that the appropriate inverse is a differentiable function. An effective algorithm for computing the derivative of the projection matrix that involves solving a set of Sylvester equations is given. Another class of methods considers the equations in a modified form, using a decomposition of the pseudogramians based on a contragredient transformation. Some freedom is left in making an exact match between the number of equations and the number of unknowns, thus effectively generating a family of methods. Three strategies are considered for balancing the number of equations and unknowns. This approach proved to be very successful on a number of examples. The tests have shown that using the ‘best’ method practically always leads to a solution. / Master of Science

在行列設計(row-column design)的架構下，當行總和與列總和皆為總試驗處理數的倍數時，我們考慮行效果與列效果的相互比較之最適性。延續Shah和Sinha(1993)的結果，在給定行總和及列總和的情況下，我們導出達成齊一最適設計(uniformly optimal design)的充分條件。此外，當總實驗單位固定時，達成全域最適設計(universally optimal design)的充分條件亦被求出。我們同時列舉許多相關的設計排列法。 / We consider the problem of comparing row effects and column effects in the row-column design setup when the row sizes and column sizes are all multiples of the number of treatments. Following the work of Shah and Sinha (1993), we derive a sufficient condition for uniformly optimal designs for given values of the row sizes and column sizes. We also derive a sufficient condition for universally optimal designs when the total number of experimental units is fixed. Several examples of designs with high efficiencies are provided.

Cheng and Wu (2001) introduced a method for response surface exploration using only one design by using a 3-level design to first screen a large number of factors and then project onto the significant factors to perform response surface exploration. Previous work generally involved selecting designs based on projection properties first and aliasing structure second. However, having good projection properties is of little concern if the correct factors cannot be identified. We apply Jones and Nachtsheim’s (2009) method for finding optimal designs with minimal aliasing to find 18, 27, and 30-run designs to use for single-step screening and optimization. Our designs have better factor screening capabilities than the designs of Cheng and Wu (2001) and Xu et al. (2004), while maintaining similar D-efficiencies and allowing all projections to fit a full second order model.

abstract: Generalized Linear Models (GLMs) are widely used for modeling responses with non-normal error distributions. When the values of the covariates in such models are controllable, finding an optimal (or at least efficient) design could greatly facilitate the work of collecting and analyzing data. In fact, many theoretical results are obtained on a case-by-case basis, while in other situations, researchers also rely heavily on computational tools for design selection. Three topics are investigated in this dissertation with each one focusing on one type of GLMs. Topic I considers GLMs with factorial effects and one continuous covariate. Factors can have interactions among each other and there is no restriction on the possible values of the continuous covariate. The locally D-optimal design structures for such models are identified and results for obtaining smaller optimal designs using orthogonal arrays (OAs) are presented. Topic II considers GLMs with multiple covariates under the assumptions that all but one covariate are bounded within specified intervals and interaction effects among those bounded covariates may also exist. An explicit formula for D-optimal designs is derived and OA-based smaller D-optimal designs for models with one or two two-factor interactions are also constructed. Topic III considers multiple-covariate logistic models. All covariates are nonnegative and there is no interaction among them. Two types of D-optimal design structures are identified and their global D-optimality is proved using the celebrated equivalence theorem. / Dissertation/Thesis / Doctoral Dissertation Statistics 2018

In this work, we introduce the Sensitivity Equation Method (SEM) as a method for approximately solving infinite dimensional optimal design problems. The SEM couples a trust-region/quasi-Newton optimization algorithm with gradient information provided by apprOXimately solving the sensitivity equation for (design) sensitivities. The sensitivity equation is (in the problems considered here) a partial differential equation (POE) which describes the influence of a design parameter on the state of the system. It is shown that obtaining design sensitivities from the sensitivity equation has advantages over finite difference and semi-analytical methods in that there is no need to remesh or compute mesh sensitivities (even if the domain is parameter dependent), the sensitivity equation is a linear POE for the sensitivities and can be approximated in an efficient manner using the same approximation scheme used to approximate the states. The applicability of the SEM to shape optimization problems, where the state is described by the Euler equations, is studied in detail. In particular, we prove convergence of the method for a one dimensional test problem. These results are used to speculate on the applicability of the method for more complex problems. Finally. we solve a two dimensional forebody simulator design problem (for use in wind tunnel experiments) using the SEM, which is shown to be a very efficient method for this problem. / Ph. D.