The hot-rolling batch scheduling method based on the prize collecting vehicle routing problem

Tao Zhang,W. Chaovalitwongse,Yue-Jie Zhang,P. Pardalos

DOI: https://doi.org/10.3934/jimo.2009.5.749

2009-01-01

Journal of Industrial and Management Optimization

Abstract:This paper studies a hot-rolling batch scheduling problem, which is a challenging problem commonly arising in the iron-steel industry. This problem deals with forming steel strips into rolling units and determining the rolling sequences while minimizing changes in characteristics (e.g., width, thickness and rigidity) of all neighbor steel strips and maximizing the machine utilization. Based on technical rules used in iron-steel production practice, we formulate this problem as a prize collecting vehicle routing problem, which is a hard combinatorial multi-objective optimization problem. We develop a new heuristic approach to solve this problem by enhancing the framework of particle swarm optimization (PSO). The key features of our approach are the utilization of number mapping function, which allows the PSO algorithm to deal with discrete problems, and the employment of tabu search at the end of every PSO iteration. We investigate and measure the performance of the proposed approach using three real life datasets obtained from a well-known iron-steel production company in China. The results suggest that our approach is very efficient and effective in providing high-quality and practical solutions, and out approach outperforms traditional PSO and tabu search algorithms based on these datasets.