采用大语言模型和Petri网的智能车间自动规划方法

打开文本图片集

中图分类号：TP273.5 文献标志码：A 文章编号： 1000-5013（2026）01-0011-09

Abstract：To enhance the human-machine interaction friendliness of engineers working with complex workshop systems，an automatic modeling and planning method for manufacturing workshops based on large language model and Petri net is proposed. First，a large language model is introduced to transform natural language descriptions of workshop operations into planning domain definition language （PDDL） specifications. Then，the corresponding Petri net model is generated by a PDDL conversion algorithm. Finally，based on the generated Petri net architecture，an optimal path search algorithm is applied to compute the optimal execution path for workshop operations，thereby realizing the automatic modeling and planning. Experimental results show that the proposed method enables efective control of machining workshop operations directly through natural language，significantly enhancing the interaction efficiency and user-friendliness for engineers in complex workshop environments.

Keywords：large language model； planning domain definition language;Petri net； automatic planning

如何使自动化工程师能够直接通过自然语言快速建模复杂的多机器人柔性制造系统1」，并获得可执行的最优控制路径，一直是工业自动化领域长期存在且极具挑战性的课题。（剩余9856字）