Probabilistic tolerance synthesis: A comparative study

The quality loss function approach advocated by Taguchi has received attention in the practice of robust design of products. The integer programming approach can be used for the selection of tolerances and manufacturing processes. The Motorola quality engineering program has been widely recognized as it reduces the number of defects by the six sigma tolerance design practice. In this paper, three approaches, namely, the Taguchi quality loss function approach, the integer programming approach, and the six sigma quality design approach are applied to probabilistic tolerance design to reduce the manufacturing variation and cost. A classical example of nonlinear tolerance design is used to illustrate the three approaches. A comparative study of the three approaches is conducted, and their advantages and disadvantages are explored.