My PhD research focuses on the recommendation of optimal designs for two-colour microarray experiments. Two-colour microarrays are a technology used to investigate the behaviour of many thousands of genes in a single experiment. This technology has created the potential for making significant advances in the field of bioinformatics. Careful statistical design is crucial to realize the full potential of microarray technology. My research has focused on the recommendation of designs that are optimal in terms of precision for effects that are of scientific interest, making the most effective use of available resources. Based on statistical efficiency, the optimality criterion used is Pareto optimality. A design is defined to be Pareto optimal if there is no other design that leads to equal or greater precision for each effect of scientific interest and strictly greater precision for at least one. My PhD thesis was submitted in June and key aspects of my research are summarised below. Pareto optimality enables the recommendation of designs that are particularly efficient for the effects that are of scientific interest. I have developed methodology to cater for effects of interest that correspond to contrasts rather than solely considering parameters of the statistical linear model. My approach also caters for additional experimental considerations such as contrasts that are of equal scientific interest. During my PhD, I have provided advice regarding the design of two-colour microarray experiments aimed at discovering the genetic basis of medical conditions. For large experiments, it is not feasible to examine all possible designs in an exhaustive search for Pareto optimal designs. I have adapted the multiple objective metaheuristic method of Pareto simulated annealing to the microarray context. The aim of Pareto simulated annealing is to generate an approximation to the set of Pareto optimal designs in a relatively short time. At each iteration, a sample of generating designs is used to explore the design space in an efficient way. This involves the setting of a number of Pareto simulated annealing parameters and the development of appropriate quality measures. I have developed algorithms to search systematically for the optimal values of the tuning parameters based on Pareto simulated annealing and response surface methodology. / Thesis (Ph.D.) -- University of Adelaide, School of Mathematical Sciences, 2010

Factorial designs have been widely used in many scientific and industrial settings, where it is important to distinguish "active'' or real factorial effects from "inactive" or noise factorial effects used to estimate residual or "error" terms. We propose a new approach to screen for active factorial effects from such experiments that utilizes the potential outcomes framework and is based on sequential posterior predictive model checks. One advantage of the proposed method lies in its ability to broaden the standard definition of active effects and to link their definition to the population of interest. Another important aspect of this approach is its conceptual connection to Fisherian randomization tests. As in the literature in design of experiments, the unreplicated case receives special attention and extensive simulation studies demonstrate the superiority of the proposed Bayesian approach over existing methods. The unreplicated case is also thoroughly explored. Extensions to three level and fractional factorial designs are discussed and illustrated using a classical seat belt example for the former and part of a stem-cell research collaborative project for the latter. / Statistics

No description available.

Choice-based conjoint experiments are used when choice alternatives can be described in terms of attributes. The objective is to infer the value that respondents attach to attribute levels. This method involves the design of profiles on the basis of attributes specified at certain levels. Respondents are presented sets of profiles and asked to select the one they consider best. Choice sets with no dominating or dominated profiles are called Pareto optimal, and these Pareto optimal choice sets are provided to respondents. However, if choice sets have too many profiles, they may be difficult to implement. Therefore, we provide strategies for reducing the number of profiles in choice sets. We consider situations where only a subset of interactions is of interest, and obtain connected main effects plans with smaller choice sets for 2^n and 3^n designs that are capable of estimating subsets of interactions inclusive of one specific factor. We also provide plans for estimating all main effects and one two-way interaction for mixed level designs. Next, we examine the relationship between certain Pareto Optimal choice sets and g-designs. Finally, we obtain connected main effects plans with smaller choice sets for estimating different subsets of interactions, not inclusive of one specific factor. / Statistics

No description available.

Apresentam-se os resultados de uma investigação sobre os fatores que influem no comportamento de revestimentos asfálticos em relação à deformação permanente, avaliado através dos ensaios de fluência por compressão uniaxial estática e dinâmica, realizados em corpos-de-prova moldados segundo as normas do ensaio Marshall. Os fatores selecionados para estudo foram o tipo de asfalto, a distribuição granulométrica do agregado, o teor de asfalto utilizado na mistura e a tensão utilizada para realização dos ensaios de fluência. Foram testados três tipos de asfalto, CAP-20 convencional, CAP-20 modificado com 4,5% de polímero SBS e CAP-20 modificado com 20% de borracha reciclada de pneu, utilizados em quatro teores. A definição dos teores asfálticos utilizados foi realizada a partir de seis dosagens Marshall iniciais, correspondentes às combinações dos três tipos de asfalto e das duas distribuições granulométricas adotadas, referentes aos centros da faixas B e C do DNER-ES 313/97. Os ensaios de fluência foram realizados sob três diferentes tensões, 0,1 MPa, 0,4 MPa e 0,7 MPa. Para avaliação dos efeitos das variações dos fatores citados no comportamento das misturas foi empregada a técnica de planejamento e análise de experimentos fatoriais fracionários assimétricos. Modelos estatísticos, referentes ao comportamento das misturas asfálticas avaliadas com os ensaios de fluência estática e dinâmica, foram obtidos com razoáveis coeficientes de determinação. / This work presents research results about influential factors in the permanent deformation behavior of asphalt mixes, evaluated through static and dynamic uniaxial creep tests, accomplished in specimens prepared according to the Marshall mix design method. The selected factors used in the analysis are type of binder, aggregate gradation, binder content and stress level used in the creep tests. Three types of binder were tested, convencional CAP-20, CAP-20 modified with 4,5% of SBS polymer and CAP-20 modified with 20% of recycled tire rubber, used in four binder content. The definition of binder content was accomplished starting from six Marshall method, corresponding to the combinations of the three type of binder and of the two adopted gradations, regarding the centers of the gradations B and C of DNER-ES 313/97. The creep tests were accomplished under three different stress, 0,1 MPa, 0,4 MPa and 0,7 MPa. Design and analysis techniques of fractional factorial experiments were used to evaluate the effects of the variations of the above mentioned factors on the mixture behavior. Statistical models of behavior were obtained with reasonable determination coefficients from the static and dynamic creep test.

Apresentam-se os resultados de uma investigação sobre os fatores que influem no comportamento de revestimentos asfálticos em relação à deformação permanente, avaliado através dos ensaios de fluência por compressão uniaxial estática e dinâmica, realizados em corpos-de-prova moldados segundo as normas do ensaio Marshall. Os fatores selecionados para estudo foram o tipo de asfalto, a distribuição granulométrica do agregado, o teor de asfalto utilizado na mistura e a tensão utilizada para realização dos ensaios de fluência. Foram testados três tipos de asfalto, CAP-20 convencional, CAP-20 modificado com 4,5% de polímero SBS e CAP-20 modificado com 20% de borracha reciclada de pneu, utilizados em quatro teores. A definição dos teores asfálticos utilizados foi realizada a partir de seis dosagens Marshall iniciais, correspondentes às combinações dos três tipos de asfalto e das duas distribuições granulométricas adotadas, referentes aos centros da faixas B e C do DNER-ES 313/97. Os ensaios de fluência foram realizados sob três diferentes tensões, 0,1 MPa, 0,4 MPa e 0,7 MPa. Para avaliação dos efeitos das variações dos fatores citados no comportamento das misturas foi empregada a técnica de planejamento e análise de experimentos fatoriais fracionários assimétricos. Modelos estatísticos, referentes ao comportamento das misturas asfálticas avaliadas com os ensaios de fluência estática e dinâmica, foram obtidos com razoáveis coeficientes de determinação. / This work presents research results about influential factors in the permanent deformation behavior of asphalt mixes, evaluated through static and dynamic uniaxial creep tests, accomplished in specimens prepared according to the Marshall mix design method. The selected factors used in the analysis are type of binder, aggregate gradation, binder content and stress level used in the creep tests. Three types of binder were tested, convencional CAP-20, CAP-20 modified with 4,5% of SBS polymer and CAP-20 modified with 20% of recycled tire rubber, used in four binder content. The definition of binder content was accomplished starting from six Marshall method, corresponding to the combinations of the three type of binder and of the two adopted gradations, regarding the centers of the gradations B and C of DNER-ES 313/97. The creep tests were accomplished under three different stress, 0,1 MPa, 0,4 MPa and 0,7 MPa. Design and analysis techniques of fractional factorial experiments were used to evaluate the effects of the variations of the above mentioned factors on the mixture behavior. Statistical models of behavior were obtained with reasonable determination coefficients from the static and dynamic creep test.