Variable neighborhood search based multi-objective dynamic crane scheduling

The paper researches a scheduling method based on variable neighborhood search (VNS) for a dynamic crane scheduling problem abstracted from steel making process in the iron and steel enterprise with random torpedo car arrivals and empty ladle breakdowns. Our objective is to minimize both makespan and total tardiness. At any rescheduling point, weights derived from artificial neural network (ANN), proper parameters for VNS are calculated that significantly enhances the performance of the scheduling method. Computational experiments indicate that the proposed method is better than those of common heuristic rules.