Abstract:Response surface approach is effective for robust parameter design. Previous response surface methodology assumes that the independent variables are measured without errors. However, this assumption might be violated due to the low capability of measurement system. This paper is concerned with applying response surface method for robust parameter design when there are measurement errors in variables. We present an unbiased estimator when there are some measurement errors and an optimal setting which is determined to minimize the expected quadratic loss. An example is illustrated to verify the effectiveness of the proposed approach. The results show that the proposed method can achieve better operating conditions under situations with measurement errors than the conventional method.

Robust Parameter Design Based on Response Surface Model under Considering Measurement Errors