Study of Fuzzy Robust Design with a
Larger-the-Better or Smaller-the-Better Characteristic
Liu Chuntao, Lin Zhihang
(State Key Laboratory for Manufacturing Systems Engineering, Xi'an Jiaotong University,
Xi'an 710049, China)
Abstract: The rules in percentile formulation for fuzzy robust design are proposed when
the product quality possesses a larger-the-better or smaller-the-better characteristic.
Based on the proposed rules, the optimal model of fuzzy robust design with the above
mentioned characteristic is presented. And the percentile search algorithm with most
probable point is adopted to evaluate the percentile of the product characteristic instead
of random simulation based methods, which is combined with the general optimal algorithm
to develop a framework for solving the presented model and improving the efficiency and
flexibility. A practical case for a solar powered irrigation system indicates that the
proposed approach enables to deal with the fuzzy robust design problems where the
high-quality probability and feasibility of design solutions can be adjusted flexibly
according to the designer¡¯s preference. Compared with the traditional fuzzy robust
design methods, the average number of function evaluations at each iterative point is
reduced by a factor of ten to greatly improve the efficiency for solving fuzzy robust
design problems.
Keywords: larger-the-better characteristic; smaller-the-better characteristic; fuzzy
robust design; percentile