An improved genetic algorithm for flexible job shop scheduling problem considering reconfigurable machine tools with limited auxiliary modules

Reconfigurable manufacturing system is widely regarded as a major drive towards the next-generation manufacturing, where one of the most common scenarios is that advanced reconfigurable machine tools (RMT) are equipped with auxiliary modules (AM) to improve the production flexibility, thus rapidly responding to market demands. Production scheduling is facing great challenges in this situation, because the limited number of AMs becomes the primary difficulty for the resource allocation. This paper investigates a flexible job shop scheduling problem with machine reconfigurations (FJSP-MR) to minimize the total weighted tardiness (TWT). Firstly, a sequence-based mixed integer linear programming (MILP) model is established for the FJSP-MR. Afterwards, considering the characteristics of the AM selection sub-problem, an improved genetic algorithm (IGA) with problem-specific encoding and decoding strategies is developed, in which an iterated local search with a modified k-insertion neighborhood structure is introduced for further optimization. Critical paths for the less common TWT objective are fully considered to ensure the local search is performed on bottlenecks of incumbent solutions. The proposed MILP model and IGA are tested on three groups of instances extended from public benchmark sets. Numerical experimental results indicate that the proposed IGA is highly efficient for the FJSP-MR in a variety of scenarios, where the specially designed local search is an effective complementary component for the basic genetic algorithm. Furthermore, a real-world FJSP-MR case is studied to show the proposed IGA is applicable to large-scale engineering problems.