An intelligent approach to generating improved product formulation for an optimal manufacturing route

Case Based Reasoning System (CBR) is an efficient and straightforward approach for recommending desirable product formulations. In response to a changeable market situation, it is vital to adapt or generate enhanced and more efficient product formulations and manufacturing routes. Due to the lack of variability and indefinite data distribution in many case libraries, however, this technique can be very limited and is inadequate for creating improved formulations. The proposed approach represents an intelligent way to overcome this shortcoming. Following a carefully premeditated Design of Experiment (DOE), a synergistic use of data noise removal, modeling and optimization techniques were utilized to develop favorable product formulations. After a multivariate validation, these along with DOE based formulations are retained in the CBR library for later use.