Intelligent factory many-objective distributed flexible job shop collaborative scheduling method

• Published in: Computers & industrial engineering 2022-02, Vol.164, p.107884, Article 107884
• Main Authors: Sang, Yanwei, Tan, Jianping
• Format: Article
• Language: English
• Subjects: Green manufacturing; Many-objective distributed flexible job shop collaborative scheduling problem (Ma-ODFJCSP); Many-objective optimization; Memetic algorithm; Production management
• ISSN: 0360-8352; 1879-0550
• DOI: 10.1016/j.cie.2021.107884

•A many-objective distributed flexible job shop collaborative scheduling model is established.•A dual-mode environment selection method is proposed.•The neighborhood structure based on collaborative adjustment of process and equipment is designed.•An effective memetic algorithm HMOMA is proposed. The many-objective distributed flexible job shop collaborative scheduling problem (Ma-ODFJCSP) is important in order to realize the green, flexible, and intelligent manufacturing process. However, as far as we know, this problem has not been considered yet in previous literature. The scheduling scale of this problem is large, and it is difficult to coordinate and optimize the scheduling of each workshop. The existing optimization algorithms cannot consider simultaneously convergence and diversity when solving the Ma-ODFJCSP. To solve this problem, firstly, we establish a many-objective distributed flexible job shop collaborative scheduling model. The scheduling objectives of the scheduling model simultaneously optimize the economic indicators and green indicators. To solve the model effectively, a high-dimensional many-objective memetic algorithm (HMOMA) is proposed. This method combines the improved NSGA-III and local search method. To effectively expand the solution set space, the neighborhood structure based on collaborative adjustment of process and equipment and based on the critical path are designed. To enhance the comprehensive performance of the population, a dual-mode environment selection method is proposed. The feasibility and competitiveness of the scheduling model and HMOMA are verified by experiment. The solution of this problem has important academic significance and engineering value for the intelligent factory.