| 摘要 |
A method is disclosed for determining a regression equation for modeling a production apparatus which processes a plurality of items in lots of different sizes. The regression equation is used to estimate a lot processing time of each of the lots. First, data representing the lot processing times is collected from the apparatus and inputted to a computer. Then, a frequency distribution for the lot processing times is determined in connection with each of the lots. Next, a plurality of sets of lot processing times are extracted from each of the frequency distributions by varying the data extraction rate. An average processing time of each of the plurality of sets of lot processing times is calculated. Then, a plurality of regression lines for the average lot processing times are determined for each of the lots. A degree of departure of each of the average lot processing times is determined based on the regression lines. Finally, the regression equation is selected from the regression lines based on the degree of departure. |
