2. What generic and specific methods are available to make a process more reliable?

2.1 Taguchi experimental designs for process design and development

When researching the operating conditions of a process, using the statistical method of experimental design optimizes the information gathered in relation to the number of experiments. They allow for a considerable reduction in the number of tests and rapid, unequivocal interpretation, provide precise results that are easy to present to non-specialists, help to study a very large number of factors, facilitate the detection of possible interactions and make modeling easier.

A particular category of experimental designs, known as "Taguchi designs", are well suited to the industrial environment, as they are easier to implement by non-specialists. What's more, they incorporate a "loss function"...