This paper presents several conceptual and algorithmic improvements to the multi-item, single-machine dynamic scheduling problem and its solution as recently presented by Jodlbauer. The dynamic demand which distinguishes the model from the majority of earlier approaches can be modified without changing the position of the optimum, but simplifying the model significantly. Furthermore, one can consistently change over to using cumulated demand in all calculations, thus eliminating several difficulties with discrete demand data. For generalization, sequence-dependent setups can be introduced. By changing from a start algorithm based on the special case of the Economic Production Lot (EPL) to one based on prioritizing product types depending on the value holding cost multiplied by capacity, the quality of the initial solution can be improved considerably. Comparison with various scheduling problems from the literature yielded a favorable outcome. The improvements in model and solution approach were found to have immediate effects on solution quality.
|Title of host publication||Proceedings FH Science Day 2006|
|Publication status||Published - 2006|
|Event||FH Science Day 2006 - Hagenberg, Austria|
Duration: 25 Oct 2006 → 25 Oct 2006
|Conference||FH Science Day 2006|
|Period||25.10.2006 → 25.10.2006|