Memetic algorithm based on learning and decomposition for multiobjective flexible job shop scheduling considering human factors

Human factors, including worker flexibility and learning–forgetting effects, are crucial factors in modern manufacturing systems to reduce costs and improve efficiency. However, traditional flexible job shop scheduling problem (FJSP) only considers machine flexibility and ignores human factors. Therefore, this paper originally investigates a multi-objective FJSP considering human factors (MO-FJSPHF) to simultaneously minimize makespan, maximum machine workload, and total machine workload. Firstly, a multi-objective mixed-integer nonlinear programming (MINLP) model is established based on the characteristics of the MO-FJSPHF. Then, a multi-objective memetic algorithm based on learning and decomposition (MOMA-LD) is proposed to solve the model by incorporating the learning-based adaptive local search into the multi-objective evolutionary algorithm based on decomposition (MOEA/D). In MOMA-LD, the machine learning technique determines which solutions deserve to perform the local search. Meanwhile, the computational resources are allocated dynamically based on the degree of population convergency during the evolutionary process. Experimental results show that our proposed algorithm outperforms four state-of-the-art algorithms on forty-three test instances and three real-world cases from a casting workshop. The validity of the proposed MINLP model is examined by the exact solver Gurobi.