ROBUST DESIGN FOR QUALITY-RELIABILITY VIA FUZZY PROBABILITY

摘要

The purpose of this paper is to describe a robust design method for quality reliability with fuzzy design target. Fuzzy number was used to express the design target of product's quality with fuzzy character, and its membership functions in common use were given. The robustness of a design solution was studied using the theory of fuzzy probability. To maximize the fuzzy probability of quality design was proposed as a robust design criterion, and its validity and practicability were validated. The controlling effect of this criterion on the mean and variance of quality index is similar to the effect of Taguchi's signal to noise ratio. This criterion is well adaptable to the different choice of the membership functions, and can effectively controls the direction at which the mean on quality index approaches to the nominal target. A robust optimal design example of a circuit was presented, and it shows that its solution is better than traditional robust design methods.