To further enhance the performance of job completion time prediction and internal due date assignment in a wafer fab, a fuzzy-neural knowledge-based system is constructed in this study. In the constructed system, multiple experts construct their own fuzzy multiple linear regression models from various viewpoints to predict the completion/cycle time of a job. Each fuzzy multiple linear regression model can be converted into an equivalent non-linear programming problem to be solved. Subsequently, a two-step aggregation mechanism is applied. At the first step, fuzzy intersection is applied to aggregate the fuzzy completion time forecasts into a polygon-shaped fuzzy number, in order to improve the precision of completion time forecasting. The polygon-shaped fuzzy number contains the actual value, and its upper bound determines the internal due date of the job. After that, a back propagation network is constructed to defuzzify the polygon-shaped fuzzy number and to generate a representative/crisp value, so as to enhance the accuracy. A practical example is used to evaluate the effectiveness of the proposed methodology. According to experimental results, the proposed methodology improved both the precision and accuracy of job cycle time prediction by 16 and 21%, respectively.
- Internal due date
- Wafer fab