Showing papers in "International Journal of Experimental Design and Process Optimisation in 2012"

Statistical analysis of some mechanical properties of friction stir welded aluminium matrix composite

Abstract: In this study, the effect of friction stir welding (FSW) parameters such as rotational speed, welding (transverse) speed, and the type of pin profile tool on some mechanical properties was statistically investigated. Plates of aluminium matrix composites fabricated by stir casting method were joined by friction stir welding process. The statistical analysis has shown that the most important factor affecting hardness and tensile strength is the welding (transverse) speed, while the rotational speed has a second ranking and pin profile tool geometry is the least. The rotational speed has no statistical significant influence on the wear rate. However, the nugget zone, which was welded by square pin profile tool, seemed to exhibit better mechanical properties compared to those obtained by other pin profile tools.

Optimally designing experiments under non-standard experimental situations

Abstract: There are numerous situations in which experiments need to be optimally designed since classic experimental design techniques are no longer effective. The experimental design space may be constrained, or already-performed experiments may have to be included. The experiment may involve qualitative factors with more than two levels, mixture and process factors in the same design, or a specific set of design points. In addition, the situation may call for reducing the number of experimental runs or using a reduced regression model in fitting the data. Finally, the region where the model is to be fitted may not be the same as where the measurements are to be made, or the model errors may have a known correlation matrix. Unfortunately, researchers in the experimental design community have paid full attention to only a few popular optimal designs; however, there are a number of other optimal designs which may be useful under different experimental situations. The main objective of this paper is the collective documentation of a broad spectrum of optimal design families, application areas, and key mathematical properties.

A research and development effort in developing the optimal formulations for new tablet drugs using design of experiments: part 1 (dissolution comparisons)

Abstract: Growth in clinical or market demand for tablet drugs often provides the impetus for increasing the scale of production. Pharmaceutical formulation optimisation is conducted initially to find the optimal combination of inactive ingredients, but changes of formulations may occur as consequence of scale-up. In this case, in vitro dissolution comparisons may need to be performed so as to demonstrate the equivalent safety and efficacy of pre-change and post-change formulations. Therefore, formulation optimisation is necessary to determine the levels of composition aimed at ensuring the equivalent safety and efficacy for the changed formulation, while meeting all related regulatory constraints. This article is an attempt to provide a comprehensive formulation optimisation procedure incorporating design of experiments techniques under different levels of scale-up changes in ingredients and biopharmaceutics classifications. Various FDA regulations and evaluation methods are investigated in order to model the optimisation scheme. Numerical examples are given in order to investigate the feasibility of the proposed methodology in solving the formulation optimisation problem for scale-up changes in composition.

Generating blocked fractional factorial split-plot designs using integer programming

Abstract: Split-plot designs are commonly used in industrial experiments when there are hard-to-change and easy-to-change factors. Due to the number of factors and resource limitations, it is more practical to run a fractional factorial split-plot (FFSP) design. These designs are variations of the fractional factorial (FF) design, with the restricted randomisation structure to account for the whole plots and subplots. When all the experimental runs cannot be performed under the same conditions, the designs are split into blocks. We discuss the formulation of blocked fractional factorial split-plot (BFFSP) designs using integer programming (IP) to achieve various design criteria.

A research and development effort in developing the optimal formulations for new tablet drugs using design of experiments: part 2 (bioequivalence studies)

Abstract: As an extension to the dissolution comparison studies in Part 1, this paper further examines the optimal levels of inactive ingredients of a formulation, when excipient changes occur which require crossover designs be performed to demonstrate bioequivalence between the pre-change and post-change formulations. In this paper, a standard 2 × 2 crossover study, as a special type of experimental design, is integrated into the ordinary framework. In addition to the discrete computational method for assessing bioequivalence that is suggested by the FDA, we further develop a continuous method to calculate critical quality characteristics. Incorporating the Taguchi quality loss concept and regulatory requirements, we propose an enhanced formulation optimisation model by simultaneously considering both the mean and variance of related characteristics. A numerical example is given to verify the effectiveness of our proposed approach. Additionally, the simulated results show that the continuous assessment method produces more desirable optimal formulations than the discrete one.

Prediction variance performance of combined array designs

Abstract: Many experiments involve variables that can be easily controlled and variables that are difficult to control (noise). In robust design one goal is to determine the settings of the controllable factors that optimise the response while simultaneously minimise the variability transmitted to the response from the noise variables. This simultaneous optimisation can be addressed using a model for the mean response and a model for the transmitted variability to fit both the control and noise variables. Combined array designs are widely used in these robust parameter design problems. We extend previous work in this area for a larger number of factors. Specifically, we investigate the prediction variance performance of a variety of combined array designs for five to 20 variables and various combinations of control and noise variables. The design region for the control variables is spherical. Prediction variance results for a number of designs and some recommendations for their use are provided.

A new joint process optimisation methodology for pharmaceutical destructive characteristics: a statistical approach

Abstract: In current pharmaceutical research and development, not many robust design (RD) and tolerance design (TD) approaches have been applied, although many researchers and practitioners have realised the importance of process design concepts. Pharmaceutical characteristics often involve different types of destructive measurements, such as hardness, friability, and disintegration in drug development and manufacturing processes. The primary objective of this paper is to develop an integrated robust-tolerance design methodology for handling destructive quality characteristics on pharmaceutical study. A statistical TD optimisation method, which incorporates consumer and producer risk, is proposed that uses a surrogate variable that is strongly correlated with the destructive quality characteristic. Finally, a pharmaceutical case study is performed for verification purposes. In the case study, a comparison between two RD optimisation models (i.e., dual-response and mean squares error models) is also conducted.

Mediation analysis with categorical independent variables

Abstract: The primary goal of mediation analysis is to explicate the mechanism that underlies an observed relationship between an independent variable and a dependent variable by including a third explanatory variable, known as a mediator variable. While the concept of mediation is theoretically appealing, our literature study indicates that there has not been a comprehensive research work on how a categorical variable is assessed in the statistical mediation analysis of time and perceived quality of life. The main objective of this paper is two-fold. First, this paper develops conceptual justifications of employing psychological variables to represent unique constructs with as little conceptual overlap as possible in the mediation analysis to prevent multicollinearity issues. Second, a modified approach is then proposed to estimate relative direct and indirect effects as well as relative total effects between a categorical independent variable and a dependent variable. A numerical example is provided to facilitate the proposed approach.

Sequential experimentation for mixtures

Abstract: We present a sequential method for experimenting with mixtures when the number of mixture components is large and the experimental goal is to gain information on a subset of components for the purpose of mixture product improvement. The model-independent method utilises parameter estimates from the Cox mixture model with a current product formulation from which to begin experimentation. The advantages of our method are the ability to perform experiments using a split-plot like structure and utilise it in a sequential manner on an operating process.