Robust Optimization: Some Tools Based in JMP® to Enhance Traditional Taguchi Methods
Aug 12, 2015 12:45 PM
simulated example transposed_20052015.jmp
case2 data set_transposed_july2015.jmp
simulated example taguchi_20052015.jmp
case2 data set taguchi_july2015.jmp
Steve Olsen, Senior Principal Reliability Engineer, Autoliv
Abstract: The idea behind Taguchi Robust Optimization is very powerful: Given the product or process we are trying to optimize, there are some factors we can’t control (or are difficult or expensive to control). Taguchi Robust Optimization methods help find the sweet spot in the design that is most robust against variation in these noise factors. Traditional Taguchi methods of experimental design and analysis, which are 50 years old, can be significantly enhanced using tools available in JMP. This paper will demonstrate how the Custom Design tool can be used to create more efficient and flexible inner and outer arrays for Taguchi DOE’s. It will show how the Profiler can automate the S/N ratio and mean graphs. We will demonstrate several methods that use the statistical analysis tools in JMP to separate signals from common cause variation – an important step which is lacking in the traditional Taguchi method. Finally, we will show how these new tools were applied to a specific problem at Autoliv and demonstrate how they yielded important insights not apparent with the traditional S/N ratio graphical analysis.
Please see the attached .pdf file for the paper. The .jmp data tables for most of the examples in the paper are also attached.