FEATURE-BASED COST ESTIMATION FOR PACKAGING PRODUCTS USING NEURAL NETWORKS

Cost estimation plays an important role in the product development cyc le. For instance, a proper cost estimation can help designers make goo d trade-off decisions regarding product structures, material, and manu facturing processes. In this paper, a feature-based product cost estim ation using back-propagation neural networks is proposed. A system usi ng this approach has been successfully developed for estimating the co st of packaging products. The cost-related features in both design and manufacturing aspects were extracted and quantified according to thei r cost drivers. The correlation between the cost-related features and the estimated costs of the product was obtained by training and valida ting a back-propagation neural network based on 60 existing products w ith their designs, process routings, and actual cost data. To illustra te, the testing results of the trained neural network based on 20 actu al products are presented. The performances of the neural network are compared to those of the company's method and a linear regression mode l. The results show that the neural network model outperformed both th e other methods in respect to performance measures such as average rel ative deviation and maximum relative deviation.