Abstract:

An integrated approach based on process capability index and entropy weight method is proposed to solve the robust parameter design for multivariate quality characteristics with individual observations. Firstly, assume that the design variables follow normal distributions and use the response surface method to build mean and variance models. Then the objective function based on process capability index is set up through analyzing the weight of deviation and variance with the entropy weight method. A numerical example shows that the optimal design variables obtained by the proposed approach could make quality characteristics satisfy the product specifications and close to the targets compared with other parameter design methods.