Control Strategies of Batch Processor with Multiple Product Types in Semiconductor Manufacturing

Abstract

Batch process machines (BPMs) process a number of jobs simultaneously as a batch. They are widely seen in semiconductor wafer fab: diffusion, oxidation and wet etch processes are examples of BPMs. Since the BPMs require long processing time and increase flow variability, they have great effect on the system performance. This thesis addresses a control problem in BPMs in wafer fab. When a BPM completes its task and is ready to start a new process, a decision should be made whether to start the process right away or to wait for the upcoming products. This thesis presents real-time control procedures for BPMs with the objective of average tardiness minimization. The proposed procedures take advantage of due-date information from current products in queue as well as upcoming products. The strategies are modeled by using a simulation software package, Arena, and the decision logic is coded in Visual Basic Application. The experimental results show that the proposed strategies give lower average tardiness compared to existing strategies.