Semiconductor product-mix estimate with dynamic weighting scheme

Semiconductor manufacturing technology is progressing rapidly. As shown by the celebrated Moor´s laws, new generations of semiconductor products emerge in the pace of every 6 months. As a result, technologies of different generations often co-exist in the same fabrication facility. For effective manufacturing resource planning, accurate prediction of product mix is therefore crucial. Product-mix prediction is usually made by performing the demand disaggregation function in demand planning. In this paper, weighted product-mix estimation methodologies are first proposed. Dynamic weighting schemes are then developed to improve the accuracy of product mix prediction. The methodologies will be tested with simulated DRAM demands and actual semiconductor demands of different technology generations.