Application of a Lagrangian relaxation based scheduling algorithm to a semiconductor testing facility

The objective of this study is the scheduling of the analog devices testing section of a production plant. The facility under study can be characterized as a high volume job shop, one of the most difficult production systems to schedule. Reentrant flow, variable daily capacity, and the need of multiple scarce resources for each operation, increase the difficulty of the scheduling problem. A Lagrangian relaxation based algorithm is proposed for scheduling the facility. The algorithm is efficient, flexible in handling the relevant modeling issues, and provides superior results when compared with simple dispatching rules which represent current industry practice.<
>