No description available.

Although there has been progress in the area of Multivariate Statistical Process Control (MSPC), there are numerous limitations as well as unanswered questions with the current techniques. MSPC charts plotting Hotelling's T2 require the normality assumption for the joint distribution among the process variables, which is not feasible in many industrial settings, hence the motivation to investigate nonparametric techniques for multivariate data in quality control. In this research, the goal will be to create a systematic distribution-free approach by extending current developments and focusing on the dimensionality reduction using Principal Component Analysis. The proposed technique is different from current approaches given that it creates a nonparametric control chart using robust simplicial depth ranks of the first and last set of principal components to improve signal detection in multivariate quality control with no distributional assumptions. The proposed technique has the advantages of ease of use and robustness in MSPC for monitoring variability and correlation shifts. By making the approach simple to use in an industrial setting, the probability of adoption is enhanced. Improved MSPC can result in a cost savings and improved quality.

This thesis consists of two topics: statistical selection and profile monitoring. Statistical selection is related to ranking and selection in simulation and profile monitoring is related to statistical process control. Ranking and selection (R&S) is to select a system with the largest or smallest performance measure among a finite number of simulated alternatives with some guarantee about correctness. Fully sequential procedures have been shown to be efficient, but their actual probabilities of correct selection tend to be higher than the nominal level, implying that they consume unnecessary observations. In the first part, we study three conservativeness sources in fully sequential indifference-zone (IZ) procedures and use experiments to quantify the impact of each source in terms of the number of observations, followed by an asymptotic analysis on the impact of the critical one. Then we propose new asymptotically valid procedures that lessen the critical conservativeness source, by mean update with or without variance update. Experimental results showed that new procedures achieved meaningful improvement on the efficiency. The second part is developing a wavelet-based distribution-free tabular CUSUM chart based on adaptive thresholding. WDFTCa is designed for rapidly detecting shifts in the mean of a high-dimensional profile whose noise components have a continuous nonsingular multivariate distribution. First computing a discrete wavelet transform of the noise vectors for randomly sampled Phase I (in-control) profiles, WDFTCa uses a matrix-regularization method to estimate the covariance matrix of the wavelet-transformed noise vectors; then those vectors are aggregated (batched) so that the nonoverlapping batch means of the wavelet-transformed noise vectors have manageable covariances. Lower and upper in-control thresholds are computed for the resulting batch means of the wavelet-transformed noise vectors using the associated marginal Cornish-Fisher expansions that have been suitably adjusted for between-component correlations. From the thresholded batch means of the wavelet-transformed noise vectors, Hotelling’s T^2-type statistics are computed to set the parameters of a CUSUM procedure. To monitor shifts in the mean profile during Phase II (regular) operation, WDFTCa computes a similar Hotelling’s T^2-type statistic from successive thresholded batch means of the wavelet-transformed noise vectors using the in-control thresholds; then WDFTCa applies the CUSUM procedure to the resulting T^2-type statistics. Experimentation with several normal and nonnormal test processes revealed that WDFTCa outperformed existing nonadaptive profile-monitoring schemes.

No description available.

This study evaluated the changes which occur in manufacturing organizations in the plastic molding industry which implement statistical process control (SPC). The study evaluated changes in product quality, consistency, cost, changes in employee attitudes, and changes in the organization structure which occur after the implementation of SPC. The study was conducted in two phases. Phase 1 consisted of an exploratory field study of a single manufacturing company. Phase 2 consisted of a field survey of three manufacturing companies in the same industry. An unexpected opportunity to evaluate the differences in effects of successful and unsuccessful SPC implementations occurred during the field survey. One plant, whose management assessed their SPC program as being unsuccessful, reported no economic or quality benefits from SPC. Neither did this plant report any changes in the attitudes or behavior of their employees. Neither of these findings was surprising since this plant was the only one of the four study plants which implemented SPC as a quality control program with no participation from the production department. The three plants whose management assessed their SPC programs as being successful reported reduced product variation and a decrease in the proportion of defective product produced as a result of SPC. No consistent evidence was found concerning a reduction in the material required per product resulting from SPC. No consistent evidence was found linking changes in employee attitudes and behavior to the implementation of SPC. The field study found a significant change in the employees1 attitudes toward management but no change in their attitudes toward the company. The field survey found no evidence of change in either dimension. Evidence was found for a change to a more organic structure during SPC training and to a more mechanistic structure during SPC implementation. The final form of the organization was more organic than before SPC. No consistent evidence was found for an increase in hierarchical levels or a reduction in the spans of control within the production departments as a result of SPC. All three of the plants with successful SPC programs increased the size and authority of their technical support staffs as a result of implementing SPC.

Healthcare is facing great challenges and psychiatric care is no exception. Extensive attempts to improve quality are made. It is essential to use methods that enable learning from experience, to improve performance. The core feature of Statistical Process Control (SPC), the control charts, are in use in various settings to enable learning and to support quality improvement work, but its use in psychiatric settings are scarce. This master´s thesis explores the usefulness of control charts, in quality improvement work. This was done in a case study at a department of psychosis by addressing two questions related to: a) control chart’s contribution to knowledge on if, when, where and how changes occur, and 2) how usefulness of control charts is perceived at the department. Control charts were applied to important variables and development officer’s and manager’s thoughts on usefulness were analysed using pattern matching. The use of charts shows shifts and differences between wards related to ongoing improvement projects. There is a readiness to start using control charts. The perceived usefulness matches the benefits and challenges identified in literature. Control charts as a tool supporting continuous improvement work in a psychiatric context, has a great potential still awaiting its use.

Essa tese versa sobre um procedimento inovador para o controle e garantia da qualidade em processos de manufatura onde existem vários atributos e variáveis a serem monitorados, como acontece, por exemplo, na indústria cerâmica. O procedimento proposto, chamado de controle integrado da qualidade, começa com a definição de alguns postos de controle na linha de produção, onde o monitoramento será feito usando uma única carta de controle que agrega vários atributos e variáveis. Esse procedimento é complementado usando gráficos de Pareto, que hierarquizam as características de qualidade que contribuem mais fortemente para o percentual de defeituosos do posto, e as cartas de controle tradicionais, que são utilizadas seletivamente, apenas conforme a indicação dos gráficos de Pareto. A utilização conjunta dessas ferramentas auxilia no diagnóstico e solução dos problemas de qualidade, permitindo que as ações de melhoria aconteçam no tempo e local certo. Entre as vantagens do procedimento proposto, cita-se: (a) a simplicidade do controle integrado, que foca prioritariamente umas poucas cartas de controle; (b) a abordagem estatística, que fornece uma base sólida para a tomada de decisões; e (c) o forte apelo gerencial fornecido pelo controle integrado, que disponibiliza a cada amostra uma visão geral da condição de qualidade nos postos de controle e na linha de produção como um todo. A tese apresenta também uma abordagem sistemática para auxiliar na implantação do controle integrado da qualidade. Essa abordagem contempla 5 etapas principais: definição do projeto, planejamento da implantação, treinamento, implantação efetiva e acompanhamento e consolidação. Por fim, é apresentado e discutido um estudo de caso em uma indústria cerâmica, que contribui para a validação do método proposto. Com base nesse estudo de caso e no referencial teórico, conclui-se sobre o assunto. / This dissertation presents a new procedure, Integrated Quality Control, for quality control and assurance in situations where severa! variables and attributes are to be monitored. That is the case in the ceramics manufacturing industry, for example. The procedure is implemented by defining control (or work) stations in the production line; variables and attributes in each station are monitored simultaneously using an aggregated control chart for the percentages of non-conformities. When a signal occurs, a Pareto graph sorts out variables and attributes identifying those contributing the most to its occurrence: such variables and attributes are then examined using traditional control charts. The joint use of aggregate charts, Pareto graphs and univariate control charts turns monitoring and improvement of processes into a non-complex task. A systematic approach for the implementation of Integrated Quality Control is also presented. The approach comprises five main steps: project definition, implementation planning, training of personnel, effective implementation, and follow up. Key advantages of the procedure suggested here are: (i) ability to handle variables and attributes in a single integrated chart, (ii) the statistical approach which provides a solid basis for the decision-making, and (iii) strong managerial appeal provided by the integrated charts. The procedure is illustrated with a case study conducted in a ceramic plant. Some conclusions are drawn based on the results of this case study and in the theoretical references.

Statistical process control (SPC) procedures suitable for autocorrelated processes have been extensively investigated in recent years. The most popular method is the residual-based control chart. To implement this method, a time series model, which is usually an autoregressive moving average (ARMA) model, of the process is required. However, the model must be estimated from data in practice and the resulting ARMA modeling errors are unavoidable. Residual-based control charts are known to be sensitive to ARMA modeling errors and often suffer from inflated false alarm rates. As an alternative, control charts can be applied directly to the autocorrelated data with widened control limits. The widened amount is determined by the autocorrelation function of the process. The alternative method, however, can not be also free from the effects of modeling errors because it relies on an accurate process model to be effective. To compare robustness to the ARMA modeling errors between the preceding two kinds of methods for control charting autocorrelated data, this dissertation investigates the sensitivity analytically. Then, two robust design procedures for residual-based control charts are developed from the result of the sensitivity analysis. The first approach for robust design uses the worst-case (maximum) variance of a chart statistic to guarantee the initial specification of control charts. The second robust design method uses the expected variance of the chart statistic. The resulting control limits are widened by an amount that depends on the variance of chart statistic - maximum or expected - as a function of (among other things) the parameter estimation error covariances.

Essa tese versa sobre um procedimento inovador para o controle e garantia da qualidade em processos de manufatura onde existem vários atributos e variáveis a serem monitorados, como acontece, por exemplo, na indústria cerâmica. O procedimento proposto, chamado de controle integrado da qualidade, começa com a definição de alguns postos de controle na linha de produção, onde o monitoramento será feito usando uma única carta de controle que agrega vários atributos e variáveis. Esse procedimento é complementado usando gráficos de Pareto, que hierarquizam as características de qualidade que contribuem mais fortemente para o percentual de defeituosos do posto, e as cartas de controle tradicionais, que são utilizadas seletivamente, apenas conforme a indicação dos gráficos de Pareto. A utilização conjunta dessas ferramentas auxilia no diagnóstico e solução dos problemas de qualidade, permitindo que as ações de melhoria aconteçam no tempo e local certo. Entre as vantagens do procedimento proposto, cita-se: (a) a simplicidade do controle integrado, que foca prioritariamente umas poucas cartas de controle; (b) a abordagem estatística, que fornece uma base sólida para a tomada de decisões; e (c) o forte apelo gerencial fornecido pelo controle integrado, que disponibiliza a cada amostra uma visão geral da condição de qualidade nos postos de controle e na linha de produção como um todo. A tese apresenta também uma abordagem sistemática para auxiliar na implantação do controle integrado da qualidade. Essa abordagem contempla 5 etapas principais: definição do projeto, planejamento da implantação, treinamento, implantação efetiva e acompanhamento e consolidação. Por fim, é apresentado e discutido um estudo de caso em uma indústria cerâmica, que contribui para a validação do método proposto. Com base nesse estudo de caso e no referencial teórico, conclui-se sobre o assunto. / This dissertation presents a new procedure, Integrated Quality Control, for quality control and assurance in situations where severa! variables and attributes are to be monitored. That is the case in the ceramics manufacturing industry, for example. The procedure is implemented by defining control (or work) stations in the production line; variables and attributes in each station are monitored simultaneously using an aggregated control chart for the percentages of non-conformities. When a signal occurs, a Pareto graph sorts out variables and attributes identifying those contributing the most to its occurrence: such variables and attributes are then examined using traditional control charts. The joint use of aggregate charts, Pareto graphs and univariate control charts turns monitoring and improvement of processes into a non-complex task. A systematic approach for the implementation of Integrated Quality Control is also presented. The approach comprises five main steps: project definition, implementation planning, training of personnel, effective implementation, and follow up. Key advantages of the procedure suggested here are: (i) ability to handle variables and attributes in a single integrated chart, (ii) the statistical approach which provides a solid basis for the decision-making, and (iii) strong managerial appeal provided by the integrated charts. The procedure is illustrated with a case study conducted in a ceramic plant. Some conclusions are drawn based on the results of this case study and in the theoretical references.

Made available in DSpace on 2014-06-11T19:35:41Z (GMT). No. of bitstreams: 0 Previous issue date: 2008-08-27Bitstream added on 2014-06-13T19:25:05Z : No. of bitstreams: 1 claro_fae_dr_guara.pdf: 832858 bytes, checksum: c3de712a6a76ca96ccda8bf5752d87f4 (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Os gráficos de X são apresentados na literatura supondo quase sempre que as observações da variável X são independentes. Na prática, no entanto, está se tornando rotina descobrir que esta condição não existe. A dependência entre observações gera um aumento na freqüência de alarmes falsos e diminui o poder do dispositivo estatístico. Nesta tese estuda-se o gráfico de X com amostragem dupla (AD) supondo que as observações de X são descritas por modelos parcimoniosos da família ARIMA (Autoregressivo, Integrado e de Médias Móveis). As propriedades da carta foram obtidas considerando o conceito de subgrupos racionais. Para comparar o desempenho do gráfico proposto com o desempenho dos esquemas concorrentes, isto é, o gráfico de X padrão, o gráfico de X com amostra de tamanho variável (ATV) e o esquema da Média Móvel Ponderada Exponencialmente (EWMA), foi necessário obter o número médio de amostras até o sinal (NMA) para todos eles. Os resultados obtidos mostram que a autocorrelação dentro do subgrupo tem forte impacto sobre as propriedades dos gráficos. O gráfico de controle com amostragem dupla é geralmente mais eficiente do que os esquemas concorrentes na detecção de desajustes na média do processo. / The X charts are presented in the literature often assuming that the observations of the X variable are independent. In practice, however, it is becoming a routine to find out that such condition is unrealistic. The autocorrelation among the observations increases the false alarm rate and reduces the power of the statistical device. In this thesis, we study the Double Sampling X chart (DS) assuming that the observations of X are described by parsimonious models of the ARIMA family (Autoregressive, Integrated and Moving Average). The properties of the charts were obtained considering the concept of rational subgroups. To compare the performance of the proposed chart with the performance of the competitor schemes, that is, the standard X chart, the Variable Sample Size X chart (VSS) and the Exponentially Weighted Moving Average (EWMA) chart, it was necessary to obtain the average run length (ARL) for all of them. The results show that the autocorrelation within the subgroup has strong impact on the chart properties. The Double Sampling X chart is usually more efficient than the competitor schemes in the detection of the process mean shifts.