A Hybrid Fruit Fly Optimization Algorithm for the Realistic Hybrid Flowshop Rescheduling Problem in Steelmaking Systems

In this study, we propose a hybrid fruit fly optimization algorithm (HFOA) to solve the hybrid flowshop rescheduling problem with flexible processing time in steelmaking casting systems. First, machine breakdown and processing variation disruptions are considered simultaneously in the rescheduling problem. Second, each solution is represented by a fruit fly with a well-designed solution representation. Third, two novel decoding heuristics considering the problem characteristics, which can significantly improve the solution quality, are developed. Several routing and scheduling neighborhood structures are proposed to balance the exploration and exploitation abilities. Finally, we propose an effective HFOA with well-designed smell and vision search procedures. In addition, an iterated greedy (IG) local search is embedded in the proposed algorithm to further enhance its exploitation ability. The proposed algorithm is tested on sets of instances generated from industrial data. Through comprehensive computational comparisons and statistical analyses, the performance of the proposed HFOA algorithm is favorably compared against several algorithms in terms of both solution quality and efficiency. Note to Practitioners-The steelmaking rescheduling process is critical to the effective operation of iron and steel production. This study models the steelmaking rescheduling problem with flexible processing time as a complex hybrid flowshop in which two types of disruptions, machine breakdown and processing variation, are considered concurrently. A weighted sum of the five objectives, including minimization of the average sojourn time, earliness penalty, tardiness penalty, cast-break penalty, and system instability penalty, is considered in the proposed algorithm. We develop an effective hybrid fruit fly optimization algorithm (HFOA) that applies two vectors to represent individuals and presents routing and scheduling neighborhood structures. An IG-based local search procedure is embedded to enhance the exploitation ability of the proposed algorithm. Two decoding heuristics considering the problem characteristics are developed. The effectiveness of the proposed HFOA is demonstrated through comparisons to other well-known and recently developed meta-heuristics. This work can be extended to practical problems by considering other types of disruptions. In addition, the proposed HFOA can also be generalized, and to other hybrid flowshop rescheduling problems.