Determining optimal targets for multi-characteristic product using fuzzy logic and Taguchi type loss function

This paper considers the targeting problem for products with multiple quality characteristics. Typically, these quality characteristics cannot be measured exactly, or related directly to the process input and/or process parameters. However, they can be estimated through an indirect relationship with the multiple observed process parameters. In practice, these relations can not be explicitly written in mathematical form. Therefore, in this paper a fuzzy relation between the observed parameters and required quality characteristics is proposed. In addition to that, a fuzzy-based process targeting model is developed, and the cost function is defined using unsymmetrical interval-based Taguchi (UIT) loss function. Also, an evolutionary algorithm is designed to solve the model and to obtain the optimal process targets. The utility of the proposed model and algorithm is illustrated by a realistic example form the literature.