Bi-Level Iterative Optimization for Integrated Scheduling of Continuous Casting-Hot Rolling

The integrated scheduling for both continuous casting and hot rolling process (CC-HR) significantly affects productivity, product completion time and customer satisfaction of steel industry. Considering the parallel consistency with the actual production, the problem is formulated as a bi-level iterative optimization model in this study. The upper-level is a mixed-integer programming model with the objective function of minimizing the time difference between casting and rolling; the lower-level model is abstracted into a vehicle routing problem (VRP) with the objective function of minimizing the width, hardness and thickness jump penalty values for hot rolling batch planning (HRBP). In order to overcome the drawback that the optimization for the lower model is easy to fall into local optimality, a hybrid algorithm combining genetic algorithm (GA) with strong global search ability and improved variable neighborhood search (VNS) with strong local search ability is proposed in this study. Finally, the upper and lower models are iteratively solved to find the scheduling scheme with smaller time difference between casting and rolling while ensuring the optimality for each objective function. The experimental results demonstrate that the established bi-level iterative model and the proposed optimization algorithm are both effective, as well as having obvious advantages in reducing the slab off-line time compared with some common methods.