知的推論モデルによる品質の定量化と培養プロセスの管理

概要

論文の詳細を見る
A quality model of an industrial product is a simulator for estimating sensory evaluations from analytical data. A process model is also a simulator for determining operational process variables from state variables. These models should be established in order to obtain a product with the desired quality. Since these relationships can only be judged by human perception as linguistic IF-THEN rules, knowledge-based modeling by means of fuzzy inferencing and artificial neural networks(ANNs)are often applied in constructing these models. In this research, a fuzzy neural network(FNN) was investigated and the following procedures were carried out; (1) Quality modeling for beverages such as ginjo sake, beer, coffee, etc., (2)process modeling for the production of these bioproducts with the aim of attaining quality control, and (3) development of an interface for estimating analytical data from human perceptual data and establishing a methodology for its application to the control of bioprocess operations.
社団法人日本生物工学会の論文
2000-03-25