Cost estimation model for semiconductor hookup construction projects

Hookup construction is one of the major tasks for a semiconductor facility construction project. The data required for conducting cost estimations of the hookup construction task are often incomplete because the whole project is undertaken in a short time or in a fast-track situation. Consequently, poor cost estimation is very likely resulted. Properly predicting the cost of a hookup construction project is crucial to control the total cost of a whole facility project. Thus, this study proposes a new cost estimation method for semiconductor hookup construction by integrating three existing approaches, including component ratios, fuzzy adaptive learning control network (FALCON), and fast messy genetic algorithm (fmGA). The proposed model is applied to 27 case projects. The results show that the average accuracy rate of the cost estimations for these case projects is around 84.85%. The accuracy rate of the proposed model has increased for about 6.07%~14.83% when comparing to those calculated by the conventional averaging method, the component ratios method and regression analysis method, respectively.