Research on Multi-objective Flexible Job Shop Scheduling Problem Based on Improved Non-Dominated Sorting Genetic Algorithm Algorithm

In the contemporary field of production management, the Flexible Job Shop Scheduling Problem (FJSP) emerges as a pivotal research topic, originating from the traditional job shop scheduling problem. In FJSP, each operation of a job can be processed by different machines, and each machine is capable of handling a variety of operations. This study constructs a multi-objective scheduling model, taking into account various factors prevalent in real-world production, such as the operating time, load, and energy consumption of machines. Addressing the limited convergence of the existing NSGA2 algorithm, this research integrates a variable neighborhood search algorithm to enhance the local search capability of the original algorithm, effectively avoiding the pitfall of local optima. To resolve the issue of low-quality initial solutions due to the random initialization of the NSGA2 algorithm, this study introduces a population initialization strategy based on heuristic rules, thereby improving the quality of initial solutions and accelerating the iterative optimization process. Comparative simulations with other heuristic rules and various intelligent algorithms demonstrate the efficiency of the improved NSGA2 algorithm in finding high-quality solutions. The application of the modified NSGA2 algorithm to real-life cases not only successfully addresses the issues of magnitude and dimensions among different objectives but also yields a satisfying set of Pareto solutions, further confirming the effectiveness and practicality of the proposed algorithm and model.