An Interval-Based Metamodeling Approach to Simulate Material Handling in Semiconductor Wafer Fabs

In this paper, we propose a new efficient meta-modeling approach as a simulation platform to estimate the performance of automated material handling systems (AMHS) in a much shorter execution time. Our new mechanism is based on imprecise probabilities, in which the simulation model parameters are represented as intervals to incorporate unknown dependency relationships as total uncertainties. The interval-based metamodel provides reasonably accurate and fast estimates of the performance measures of interest. The performance measures from the interval-based simulation are represented as intervals that enclose the traditional real-valued simulation estimates. Using the SEMATECH virtual fab as a test bed, the metamodel of the wafer fab AMHS is implemented in JSim, a java-based discrete-event simulation environment, and the results are compared to the detailed large-scale simulation model to investigate the validity of the proposed approach