Flexible Capacity Planning Under Uncertain Demand

Abstract

This thesis is created to explore strategies for adaptable capacity planning uncertain demand scenarios, driven by the inefficiencies of current methods during economic downturns. Businesses have been shifting their trade-off in priority from mitigating risks to minimizing costs of operation. Initially, the feasibility of employing fuzzy numbers for demand forecasting was examined. However, due to data limitations and the absence of historical records, particularly in the research of New Product Introduction (NPI), this approach was found impractical. Through a case study of Intel Vietnam, a semiconductor assembly facility, insights were gleaned into discerning trends and recurrent patterns in previous product launches within similar product families. Subsequently, a linear regression model was devised to minimize discrepancies between current forecasts and historical data, resulting in an adjusted forecast that enhances precision compared to traditional methods. Essential factors and methodologies crucial for production capacity planning, such as efficient scheduling and procurement strategies, are studied. The implementation of the new approach yields significant enhancements in the accuracy of production capacity planning by streamlining extensive planning endeavors. Overall, the thesis aims to address challenges in production capacity planning within Intel Vietnam's production department, emphasizing the imperative of cross-departmental collaboration, particularly between forecasting and engineering teams, to foster improved communication and effectiveness in tackling these challenges.