The Taguchi method emphasizes the design of a robust product that is insensitive to disturbances (noises). This feature of the Taguchi method is useful in mechanical design since an optimal design obtained through the application of the Taguchi method will also be inherently robust. The Taguchi concept of design of experiments (DOE) is used to solve unconstrained and constrained optimization problems. The signal-to-noise ratio used in the conventional Taguchi method uses only positive function values and hence is not applicable to design optimization directly. In this work, a modified form of signal-to-noise ratio is proposed so that both positive and negative objective function values can be handled. In the case of unconstrained optimization, the search directions are determined by using the Taguchi philosophy. The partial factorial design of experiments provides an efficient way of determining a good search direction. The procedure is extended to constrained optimization problems in the presence of both single and multiple objective functions. Detailed solution procedures are presented along with three numerical examples for illustration.
