Clearing function-based simulation optimization for release planning under digital twin wafer fabs

Abstract

This research studies the generating and updating of production release plans in digital twin wafer fabs. We collect the manufacturing data of the workload and the expected output for each planning period from the manufacturing execution system and the data monitoring system. Then, we fit the above data as piecewise linear clearing functions (CFs) and import the parameters of clearing functions to the mathematical planning model. A theoretical optimal solution can be calculated by the mathematical model. We import this solution as an initial solution into the discrete-event simulation model to perform simulation iterative optimization, until a satisfactory solution is obtained. To reflect the physical production system data changes in time and to update production plans, we design two update strategies. Strategy 1: The discrete event simulation model in virtual space is updated when key parameters exceed the designed thresholds. Strategy 2: The simulated data used by clearing functions is updated after every production planning process. We fit clearing functions using both historical data and simulation prediction data.