Artificial bee colony algorithm for a single machine maintenance problem with periodic machine availability constraints

In scheduling problems, machines are often assumed to be available at all times, and maintenance and setup times are often ignored. However, this assumption is invalid for real-life problems. In fact, in a production system, machines are often either periodically available or unavailable. Machine availability is referred to in the literature as machine maintenance because it is generally assumed that these availability periods are used to perform maintenance activities. In this study, the single machine periodic maintenance problem, which has periodic machine availability constraints, is investigated. Another reason for considering this research is the requirement to complete the manufacturing operations within the shift/working hours. Therefore, in the single machine scheduling problem, it is aimed to find a completion time suitable for the availability of the periodic machine. Since this problem is NP-hard, approximate results are obtained using different heuristic algorithms. Additionally, in previous studies, this problem was compared to the classical packaging problem (bin packing), and results were obtained. In this study, the single machine periodic maintenance problem with periodic machine availability constraint is solved using Artificial Bee Colony (ABC) algorithm and the results are compared with previous studies. It has been seen that the ABC algorithm does not give good results in large problems, but is effective in small problems. For this, a New Best Fit (NEW_BF) heuristic in the literature was included in the ABC algorithm and a improved ABC algorithm was proposed. The proposed algorithm gave good results for all problem sizes. The results of the improved ABC algorithm were compared with the NEW-BF and its effectiveness was proven.