IDENTIFICATION OF ROBUST SCHEDULING HEURISTICS - APPLICATION OF TAGUCHI METHODS IN SIMULATION STUDIES

The successful implementation of scheduling heuristics developed by researchers and practitioners can be greatly influenced by the heuristic's ability to perform well across a wide spectrum of environments. This particular characteristic of a heuristic, robustness, can be quantified and analyzed via the experimental strategies and techniques introduced by Taguchi. This paper examines the concept of robust scheduling heuristics, introduces experimental designs which are useful for analyzing robustness in simulation studies, and discusses appropriate statistical analysis. A comparison of different scheduling and due date assignment rules is used for illustrative purposes. It is shown that the selection of the 'best' heuristic may change when the experimental results are looked at in terms of robustness and dispersion.