Choice of an experimental design is an important concern for most researchers. Judicious selection of an experimental design is also a weighty matter in Robust Parameter Design (RPD). RPD seeks to choose the levels of fixed controllable variables that provide insensitivity (robustness) to the variability of a process induced by uncontrollable noise variables. We use the fact that in the RPD scenario interest lies primarily with the ability of a design to estimate the noise and control by noise interaction effects in the fitted model. These effects allow for effective estimation of the process variance — an understanding of which is necessary to achieve the goals of RPD. Possible designs for use in RPD are quite numerous. Standard designs such as crossed array designs, Plackett-Burman designs, combined array factorial designs and many second order designs all vie for a place in the experimenters tool kit. New criteria are developed based on classical optimality criteria for judging various designs with respect to their performance in RPD. Many different designs are studied and compared. Several first-order and many second order designs such as the central-composite designs, Box-Behnken designs, and hybrid designs are studied and compared via our criteria. Numerous scenarios involving different models and designs are considered; results and conclusions are presented regarding which designs are preferable for use in RPD. Also, a new design rotatability entity is introduced. Optimality conditions with respect to our criteria are studied. For designs which are rotatable by our new rotatability entity, conditions are given which lead to optimality for a number of the new design comparison criteria. Finally, a sequential design-augmentation algorithm was developed and programmed on a computer. By cultivating a unique mechanism the algorithm implements a D_s-optimal strategy in selecting candidate points. D_s-optimality is likened to D-optimality on a subset of model parameters and is naturally suited to the RPD scenario. The algorithm can be used in either a sequential design-augmentation scenario or in a design-building scenario. Especially useful when a standard design does not exist to match the number of runs available to the researcher, the algorithm can be used to generate a design of the requisite size that should perform well in RPD. / Ph. D.

A new optimization technique for SPICE modeling parameters has been developed in this dissertation to increase the accuracy of the circuit simulation. The importance of having accurate circuit simulation models is to prevent the very costly redesign of an Integrated Circuit (IC). This radically new optimization technique utilizes the Taguchi method to improve the fit between measured and simulated I-V curves for GaAs MESFETs. The Taguchi method consists of developing a Signal-to-Noise Ratio (SNR) equation that will find the optimum combination of controllable signal levels in a design or process to make it robust or as insensitive to noise as possible. In this dissertation, the control factors are considered the circuit model curve fitting parameters and the noise is considered the variation in the simulated I-V curves from the measured I-V curves. This is the first known application of the Taguchi method to the optimization of IC curve fitting model parameters. In addition, this method is not technology or device dependent and can be applied to silicon devices as well. Improvements in the accuracy of the simulated I-V curve fit reaching 80% has been achieved between DC test extracted parameters and the Taguchi optimized parameters. Moreover, the computer CPU execution time of the optimization process is 96% less than a commercial optimizer utilizing the Levenberg-Marquardt algorithm (optimizing 31 FETs). This technique does a least square fit on the data comparing measured currents versus simulated currents for various combinations of SPICE parameters. The mean and standard deviation of this least squares fit is incorporated in determining the SNR, providing the best combination of parameters within the evaluated range. Furthermore, the optimum values of the parameters are found without additional simulation by fitting the response curves to a quadratic equation and finding the local maximum. This technique can easily be implemented with any simulator that utilizes simulation modeling parameters extracted from measured DC test data. In addition, two methods are evaluated to obtain the worst case modeling parameters. One method lobks at the correlation coefficients between modeling parameters and the second looks at the actual device parameters that define the +/- 3σ limits of the process. Lastly, an example is given that describes the applicability of the Taguchi methodology in the design of a differential amplifier, that accounts for the effect of offset voltage. / Ph. D.

A research report submitted to the Faculty of Engineering and the Built Environment, University of the Witwatersrand, Johannesburg, in partial fulfilment of the requirements for the degree of Master of Science in Engineering. Date: September 2017, Johannesburg / The focus in this research investigation is to investigate the Process Parameter Optimisation in Laser Beam Forming (LBF) process using the 4.4 kW Nd: YAG laser system – Rofin DY 044 to form 200 x 50 x 3 mm3 mild steel - AISI 1008 samples. The laser power P, beam diameter B, scan velocity V, number of scans N, and cooling flow C were the five input parameters of interest in the investigation because of their influence in the final formed product. Taguchi Design of Experiment (DoE) was used for the selection and combination of input parameters for LBF process. The investigation was done experimentally and computationally. Laser Beam Forming (LBF) input parameters were categorised to three different levels, low (L), medium (M), and high (H) laser forming (LBF) parameters to evaluate parameters that yield maximum bending and better surface finish/quality. The conclusion drawn from LBF process is that samples which are LBFormed using low parameter settings had unnoticeable bending and good material surface finishing. On the other hand, samples LBFormed using medium parameters yielded visible bending and non-smooth surface finishing, while samples processed using high LBF parameters yielded maximum bending and more surface roughness than the other two process parameters. / MT2018