Robust Optimization / Taguchi Method
Create failure resistant products/processes.
Robust Optimization/Taguchi Methods is a subset of the DFSS (Design for Six Sigma) Methodology. It helps engineers improve the robustness of a product or process using low-cost variations of a single, conceptual design. Robust Optimization uses Robust Assessment to estimate the robustness of low-cost combinations of design parameter (control factor) values with a signal conceptual design to discover the most robust combination of design parameter (control factor) values.
In product or process development, Robust Optimization occurs after system (conceptual) design is complete and before the conceptual design is adjusted to meet requirements. Robust Optimization as a first step discovers a robust, low-cost combination of control factors. This combination may not meet or even be close to meeting the design requirements. However, Robust Optimization as a second step discovers how to adjust that low-cost combination of control factors so that the product can meet (or exceed) requirements.
In recent years, several new powerful tools have been developed within the Robust Optimization/Taguchi Methods and include MTS (Mahalanobis Taguchi System) and SBT (System Behavior testing). MTS can reduce false negatives and false positives in diagnostic systems and can be applied to pattern recognition systems such as voice/face recognition and artificial Intelligence systems. SBT (System Behavior testing) can test large combinations of system hardware and software interfaces to determine potential failures quickly and inexpensively.