Hybrid Gaussian quantum particle swarm optimization and adaptive genetic algorithm for flexible job-shop scheduling problem

• Published in: Engineering applications of artificial intelligence Vol. 154; p. 110882
• Main Authors: Xu, Yuanxing, Zhang, Mengjian, Wang, Deguang, Yang, Ming, Liang, Chengbin
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 15.08.2025
• Subjects: Adaptive genetic algorithm; Artificial intelligence; Chaotic encoding; Flexible job-shop scheduling problem; Gaussian quantum particle swarm optimization; Flexible job-shop scheduling problem; Chaotic encoding; Gaussian quantum particle swarm optimization; Adaptive genetic algorithm; Artificial intelligence
• ISSN: 0952-1976
• DOI: 10.1016/j.engappai.2025.110882

Flexible job-shop scheduling problem (FJSP) is a critical challenge in both modern manufacturing and artificial intelligence (AI), characterized by its non-deterministic polynomial-time hard nature. Despite advancements in optimization algorithms, balancing solution quality, computational efficiency, and convergence speed remains challenging, especially for large-scale FJSP instances. Although existing methods are effective in certain scenarios, they often struggle to balance global exploration and local refinement within the solution space. This study proposes a hybrid Gaussian quantum particle swarm optimization and adaptive genetic algorithm (HGQPSO-AGA) to address these limitations. HGQPSO-AGA leverages the global search capability of Gaussian quantum particle swarm optimization (GQPSO) inspired by quantum mechanics with the local refinement strength of adaptive genetic algorithm (AGA). To further enhance performance, the algorithm employs a chaotic encoding scheme, a multi-subpopulation selection mechanism, and adaptive crossover and mutation operators. The performance of HGQPSO-AGA is evaluated on three widely used FJSP benchmark datasets (Kacem, Brandimarte, and Dauzere-Peres) and validated through an industrial case study. The results demonstrate that HGQPSO-AGA outperforms comparison algorithms significantly in terms of solution quality, computational efficiency, and convergence speed. Specifically, HGQPSO-AGA achieves average makespan reductions of 2.40, 7.84, and 110.49 for Kacem, Brandimarte, and Dauzere-Peres datasets, respectively, and 5.84 in the industrial case. These findings indicate that HGQPSO-AGA provides a robust and efficient solution to FJSP, with wide-ranging implications for optimizing complex scheduling problems in manufacturing and AI-driven applications. •HGQPSO-AGA is proposed to solve the flexible job-shop scheduling problem (FJSP).•Four chaotic encoding schemes are compared to a classical two-layer encoding scheme.•HGQPSO-AGA applies chaotic encoding, subpopulation selection, and adaptive operators.•Benchmark tests and a real case verify the effectiveness of HGQPSO-AGA for FJSP.