An Effective Artificial Bee Colony Algorithm for a Real-World Hybrid Flowshop Problem in Steelmaking Process

This paper aims to provide a solution method for the real-world hybrid flowshop scheduling problem resulting from a steelmaking process, which has important applications in modern iron and steel industry. We first present a mixed integer mathematic model based on a comprehensive investigation. Then, we develop a heuristic method and two improvement procedures for a given schedule based on the problem-specific characteristics. Finally, we propose an effective artificial bee colony (ABC) algorithm with the job-permutation-based representation for solving the scheduling problem. The proposed ABC algorithm incorporates the heuristic and improvement procedures as well as new characteristics including a neighboring solution generation method and two enhanced strategies. To evaluate the proposed algorithm, we present several adaptations of other well-known and recent metaheuristics to the problem and conduct a serial of experiments with the instances generated according to real-world production process. The results show that the proposed ABC algorithm is more effective than all other adaptations after comprehensive computational comparisons and statistical analysis. Note to Practitioners-The steelmaking process, processing hot metal to steel with a well-defined chemical composition and solidifying the steel into slabs, is usually the bottleneck in iron and steel production, and thus effective scheduling methods are crucial to improve productivity of the production system. This paper models the steelmaking scheduling problem as a complex hybrid flowshop including three successive stages (steelmaking, refining, and continuous casting) with each stage having multiple parallel machines. The objective is to minimize the earliness/tardiness penalty and the average sojourn times. We develop an effective artificial bee colony (ABC) algorithm that applies job permutations to represent individuals, and present a new neighboring solution generation method for both the employ and onlooker bees, and propose two enhanced strategies to balance the exploration and exploitation. Furthermore, a heuristic method is presented to generate a good initial solution at negligible computational effort, and two improvement procedures are provided to improve a schedule based on the problem-specific characteristics. The effectiveness of the proposed ABC is demonstrated by comparisons against the adaptations of other well-known and recent metaheuristics. Since the scheduling problem in steelmaking process is very complex, in some situations buffers capacity and/or transporting capacity are limited and have to be considered. This work can be extended to these practical problems by considering the buffers capacity and/or transporting capacity in the mathematic model. In addition, the application of the proposed ABC can also be generalized to other hybrid flowshop scheduling problems.