A Genetic Algorithm Approach for Minimizing Earliness-Tardiness on a Single Burn-In Oven with a Common Due Date

摘要

We consider the problem of minimizing earliness-tardiness on a single burn-in oven with a common due date. Problems of this type have great practical relevance in the backend area of a semiconductor wafer fabrication facility. We suggest a heuristic algorithm. In the first stage we use job sequencing rules and dynamic programming in order to form batches. Then we apply a genetic algorithm for obtaining a partition of the job set into a set of early jobs and a set of tardy jobs (stage two). We use a job sequencing rule and dynamic programming for batch formation on the two sets from second stage. We determine an optimal sequence of batches by using the batch weighted longest processing time rule and the batch shortest processing time rule. We compare the performance of the suggested algorithm with a property-based heuristic suggested by Sung et al. (2002).