Systematic generation of alternative production schedules

We propose a systematic approach for generating alternative production schedules to address two challenges in implementing optimization-based scheduling techniques in industrial settings: (a) the limited scope of the scheduling models due to modeling simplifications, and (b) the consideration of shop-floor nervousness. We first introduce metrics to quantify specific characteristics of a schedule (e.g., number of batches and degree of nervousness). Next, we favor, at different degrees, such characteristics in each alternative schedule, by penalizing the metrics in the objective function. Through illustrative instances, we show that multiple alternative schedules, including schedules with desired properties can readily be generated.