转换
拖延
计算机科学
生产(经济)
缩小
算法设计
作业车间调度
数学优化
实时计算
算法
调度(生产过程)
生产经理
优化算法
电子邮件
生产控制
自动化
序列(生物学)
遗传算法
工程类
生产线
处理器调度
线性规划
生产计划
作者
Lilong Yang,Guangyu Lu,Zhengkai Li,Hao Sun,Xinghu Yu,Zhitai Liu,Zhan Li,Juan Jose Rodriguez-Andina
标识
DOI:10.1109/tii.2026.3727548
摘要
In high-mix, low-volume surface mount technology manufacturing, production lines face the pressure of frequent changeover setup operations and the challenge of timely order completion under different due dates. To address these issues, this article proposes a methodology that promotes greater feeder sharing through printed circuit board (PCB) grouping to reduce changeover setup time, while enabling flexible sequencing through accurate production prediction. Within this methodology, a hierarchical mixed-integer programming model is first developed, aiming to minimize changeover setup time and delivery tardiness. Subsequently, a feeder-aware grouping and predictive sequencing (FAG-PS) algorithm is developed to address the formulated optimization problem through two stages: grouping and sequencing. In the grouping stage, a demand-driven feeder allocation algorithm determines the required feeders. A commonality-based constructive heuristic and an iterative clustering algorithm then sequentially generate and refine the grouping solution. In the sequencing stage, a gradient boosting model accurately estimates PCB production times. A bilevel sequencing optimization algorithm is then executed to minimize intergroup and intragroup delivery tardiness. Experimental results demonstrate that FAG-PS outperforms state-of-the-art algorithms, achieving average reductions of 6.59% in changeover setup time and 6.44% in delivery tardiness.
科研通智能强力驱动
Strongly Powered by AbleSci AI