Rescheduling of a Single Machine Problem Thanks to Time Dependent Variable Weights
Résumé
Rescheduling process consists in ordering in the time all production orders in response to disruptions for optimizing a criterion. When the disruption is due to orders arrival, the heavy weight orders fit into the schedule forcing the low weight ones to wait. Therefore, the low weight orders may not run. For dealing with this issue, in this paper, we propose to reschedule with time dependent variable weights. Indeed, the longer the task remains in the system, the more its weight increases. A single machine rescheduling problem is treated with orders arriving over time. In the predictive phase, a MILP model is implemented to solve a scheduling problem with the Total Weighted Waiting Time (TWWT) as an objective. In the reactive phase, when a new order arrives, a MILP is regenerated solving the new problem with the TWWT and the stability as an objective. Experimental results discuss the impact of increasing the jobs weight and its relationship with the system performance and the resolution time.