Real Time Quality Optimization in Flexible Manufacturing Systems

Abstract

In this paper a quality oriented model is proposed for product and process as a whole. A hierarchical structural description allows both model consistency and integration of detailled information. Material and data flows are so perfectly defined. Inaccurate or uncertain data are modelled by mean of fuzzy quantities. Associated operators and relations permit information propagation in the model. That strict but flexible approach shows actual advantages above standard statistical methods. Product quality and process operation are valued in order to build a goal cost function. That criterion is used to define the optimal control strategy. A distributed computation process reduces information exchanges and contributes greatly to a safe on-line quality management.