基于启发式Q学习的汽车涂装车间作业排序优化

针对汽车涂装车间中的作业优化排序问题,提出一种基于启发式Q学习的优化算法。首先,建立包括满足总装车间生产顺序和最小化喷枪颜色切换次数的多目标整数规划模型。将涂装作业优化排序问题抽象为马尔可夫过程,建立基于启发式Q算法的求解方法。通过具体案例,对比分析了启发式Q学习、Q学习、遗传算法三种方案的优劣。结果表明:在大规模问题域中,启发式Q学习算法具有寻优效率更高、效果更好的优势。本研究为机器学习算法在汽车涂装作业优化排序问题的应用提出了新思路。

Optimization on Car Sequencing Problem in Automotive Paint Shops with Heuristic Q-learning

Aiming at the optimization of job scheduling in automotive paint shops, an optimization algorithm based on heuristic Q learning is proposed. For starters, a multi-objective integer programming model is established with the purpose of keeping painting sequence consistent with the production sequence of the assembly shop and minimizing the number of color switches. The optimization problem of painting job is abstracted into a Markov process, and a solution method based on heuristic Q algorithm is established. Through the specific cases, the advantages and disadvantages of heuristic Q learning, Q learning and genetic algorithm are compared and analyzed. The results show that the heuristic Q learning algorithm has the advantages of higher efficiency and better effect, which are more pronounced in large-scale issues. This study proposes a new idea for the application of machine learning algorithms in the optimization of job scheduling in automotive paint shops.