Solving the steelmaking casting problem using an effective fruit fly optimisation algorithm

This paper presents an effective fruit fly optimisation algorithm (FOA) to solve the steelmaking casting problem. First, we model the realistic problem as a hybrid flow shop (HFS) scheduling problem with batching in the last stage. Next, the proposed FOA algorithm is applied to solve the realistic HFS problems. In the proposed algorithm, each solution is represented by a fruit fly. Each fruit fly first improves its status through a well-designed smell search procedure. During the vision-based search procedure, the worst fruit fly in the population will be induced by the best fruit fly found thus far to improve the exploitation ability of the entire fruit fly population further. To enhance the exploration ability of the proposed algorithm, in each generation, each fruit fly that has not updated its status during the last several iterations will be replaced by a newly-generated fruit fly. The proposed algorithm is tested on sets of the instances that are generated based on the realistic production. Moreover, the influence of the parameter setting is also investigated using the Taguchi method of the design-of-experiment (DOE) to determine the suitable values for the key parameters. The results indicate that the proposed FOA is more effective than the four presented algorithms.