Abstract
This paper describes a new approach to an adaptive lot release scheduling using genetics based machine learning (GBML) in a semiconductor manufacturing system. The proposed method employs a scheduling agent with a classifier system (CS) to release materials onto the production floor properly according to the state of the production floor. Effectiveness of the proposed GBML based lot release method is discussed by using results of the computer simulations in terms of not only productivity but also adaptability to the production demand change.
