Mini–max regret strategy for robust capacity expansion decisions in semiconductor manufacturing

Semiconductor manufacturing is capital intensive and the capacity utilization significantly affects the capital effectiveness and profitability of semiconductor manufacturing companies. Semiconductor companies have to rely on various demand forecasts of the products fabricated by different technologies as a basis to determine capital investments for capacity expansion plans that require long-lead time. Moving into the consumer electronics era, semiconductor products have increasingly shortened product life cycles with demand fluctuation. However, given the involved uncertainties and long lead-time for capacity expansion, the capacity expansion decisions of semiconductor manufacturing companies are under both the risks of capacity shortage and surplus along with the time. This study aims to propose a mini–max regret strategy for capacity planning under demand uncertainty to improve capacity utilization and capital effectiveness in semiconductor manufacturing. This model considers each of possible outcomes of the multi-period demand forecasts from sales and marketing departments to generate robust capacity expansion plan to minimize the maximum regrets of capacity oversupply and shortage. An empirical study was conducted in a leading semiconductor company in Taiwan for the validation. The results have shown robustness and practical viability of the proposed mini–max strategy for capacity expansion planning under demand uncertainties.