Rapid Unconstrained Fab Model using a Business Intelligence Tool DM: Data Management and Data Mining Tools

Semiconductor FABs operate with dynamic demand and supply changes as compared to any other industry. The industry traditionally uses spreadsheet based capacity models as well as more sophisticated tools such as Advanced Planning & Scheduling and/or Discrete Event Simulation software. While spreadsheet models are great for simple scenarios, fab managers are increasingly relying more on non-spreadsheet engines. Non-spreadsheet methods are time consuming and typically take hours to prepare data and hours to run. If the results are not acceptable, the cycle repeats. This paper attempts to present a new approach in using a state-of-the-art business intelligence (BI) tool for solving one subset of the problem which is forecasting the flow of lots in the line using an unconstrained modeling approach. Combined with the power of BI tool with a minimal backend scripting we present an approach that runs a whole fab model for many months in single digit minutes.