Solving assembly line balancing problems with reinforcement learning

doi:10.1016/j.engappai.2025.113110

Solving assembly line balancing problems with reinforcement learning

2025-11-12

0

PRE

AI

DOI：10.1016/j.engappai.2025.113110

Original

Original request for help

Save

Abstract

En 中文

This paper introduces and demonstrates the first direct application of a reinforcement learning algorithm to the standard assembly line balancing problem in literature. The proposed approach is computationally efficient and can be easily extended to other types of assembly line balancing problems. To assess its real-world applicability, the algorithm was applied to two case studies, demonstrating its effectiveness in achieving a balanced workload distribution while maintaining line efficiency. Additionally, a comprehensive comparative analysis was conducted, evaluating the performance of the reinforcement learning-based method against metaheuristic algorithms, the Computer Method of Sequencing Operations for Assembly Lines (COMSOAL) heuristic algorithm, and individual task assignment rules. The results highlight the superior capability of the proposed approach in minimizing workload variation across workstations. The statistical robustness of these findings is validated through Friedman, Wilcoxon signed-rank, paired-t, and Levene's tests. The proposed approach reliably achieved optimal solutions across a broad range of test cases, emphasizing its adaptability, reliability, and effectiveness in addressing assembly line balancing problems.

Solving assembly line balancing problems with reinforcement learning

Abstract

Journal

Organization