Project Abstract:

The problem of scheduling semiconductor wafer fabrication facilities (wafer fab) has been considered to be one of the most important scheduling problems. Of particular interest is the presence of batch processing machines (BPMs) within the wafer fab such as oxidation furnaces. A batch processing machine is defined as a piece of equipment that has capacity greater than one. Due to their long processing times and ability to act as a situational bottleneck, the manner in which BPMs are scheduled can have a very strong influence on system performance. However, little work has been done on scheduling BPMs in the context of it being part of an entire wafer fab system; majority of research on scheduling wafer fabs ignore the presence of BPMs. Although the problem of scheduling BPMs has attracted greater attention from researchers, majority have considered truncated systems involving BPMs; these generally come in the form of heuristics. This heterogeneity between the existing literature regarding scheduling wafer fabs, and scheduling BPMs, is a critical gap in the existing literature that this project proposes to address. It is anticipated that explicitly modeling the potential impact of BPMs on the quality of a particular schedule for a wafer fab will lead to the development of a feasible scheduling algorithm that improves upon industry’s current best practices.

Prior to determining the potential impact of the BPM on a complex manufacturing system is, an analytical characterization of the BPM in the presence of other serial processors is required, for several reasons. By analyzing the phenomena associated with scheduling a small system with a BPM, a deeper understanding of the relationships between the BPM and its upstream and downstream processors can be obtained. Such an understanding would then allow for the formulation of policies that can control serial processors with overall improvement of the scheduling quality for the entire system in mind. This is a marked departure from previous look-ahead methods. This shift in scheduling philosophy, from a reactive to a proactive approach, is expected to yield significant improvement in schedule quality over previous look ahead methods.