In a system-level design, the performance of that system (output) can be affected by the variation of input values. Such input variations can be caused by component tolerances and / or changes in the usage environment, all of which are a fact of life. Therefore, it’s critical that the Designer understands the sensitivity of the system performance across the potential spectrum of input parameters. In systems comprised of numerous individual components, the sheer number of permutations can quickly become unmanageable.
Design of Experiments (DoE) is a statistical technique that systematically determines which inputs have a significant impact on the output in a cost effective and time efficient manner. The process designs a set of experiments (model runs of different configurations), which results in a response surface that identifies which variables have the most influence on the outcome. This can then guide the design to a less sensitive layout or it can guide tolerance selection to maintain acceptable performance.
Download more information here: Pushing the Boundaries of Experimental Design with Design of Experiments (DOE)