نوع مقاله : پژوهشی
تازه های تحقیق
عنوان مقاله English
نویسنده English
This study introduces an innovative approach to addressing the dual challenges of line balancing and sequencing in mixed-model assembly lines through the integration of artificial intelligence. By combining deep neural networks (DNNs) and reinforcement learning (RL), the proposed framework simultaneously optimizes workstation load distribution and task ordering. Historical production data—including task processing times, the total number of workstations, and inter-task dependencies—were collected and structured into predictive models. A Deep Q-Network (DQN)-based RL agent was developed to dynamically assign tasks to stations and determine their execution sequence in real time, aiming to minimize makespan and maximize overall line efficiency. In parallel, DNNs were employed to forecast task processing durations and evaluate the feasibility of relocating tasks across stations. Numerical experiments using real-world production data demonstrate that the proposed method significantly reduces idle time, decreases task waiting periods, and streamlines workflow continuity. Furthermore, benchmarking against conventional optimization techniques—such as Genetic Algorithms and Simulated Annealing—highlights the advantages of this machine learning–driven strategy, particularly in achieving near-optimal solutions more rapidly and with greater adaptability.
کلیدواژهها English
Copyright © Fahimeh Tanhaie
License
This article is released under the Creative Commons Attribution (CC BY 4.0) license. Anyone is free to copy, share, translate, and adapt this article for any purpose, whether commercial or non-commercial, as long as proper citation is given to the authors and original publication.
