自动化学报2026，Vol.52Issue(1)：78-90,13.DOI:10.16383/j.aas.c250308
基于多智能体强化学习的流程工业多操作参数协同优化
Collaborative Optimization of Multiple Operating Parameters for Process Industries Based on Multi-Agent Reinforcement Learning
摘要
Abstract
Process industries are often confronted with strong multi-operational parameter couplings,intricate pro-cess topologies,and difficulties in multi-stage coordination,which render conventional localized optimization meth-ods inadequate for achieving global optimality.To address these challenges,this paper proposes a graph spectral theory-based process topology-aware multi-agent reinforcement learning collaborative optimization method for mul-tiple operating parameter collaborative optimization in complex topological process industries.Specifically,a topo-logy analysis framework based on Laplacian spectral analysis is developed to characterize structural coupling rela-tionships among multiple operating parameters,thereby supporting agent task allocation and coordinated decision-making.Subsequently,a temporal perception module integrating long short-term memory networks with a multi-head attention mechanism is designed to extract key temporal dependencies from historical state trajectories.Fur-thermore,a hierarchical spatial attention mechanism is introduced to enable dynamic and adaptive regulation of op-timization attention across organizational,variable,and continuous control domains.On this basis,a hierarchical reinforcement learning architecture is constructed to coordinate local and global policy optimization,facilitating co-operative control and strategy optimization among multiple agents.Simulation experiments using industrial data from a continuous stirred tank reactor system and a representative salt-lake chemical process validate the effective-ness of the proposed method.Experimental results show that the proposed method achieves up to a 41.2%perform-ance improvement over conventional approaches,exhibiting superior convergence behavior and policy stability,and providing a viable technical pathway for multiple operating parameter collaborative optimization in process industries.关键词
协同优化/图谱感知强化学习/拉普拉斯谱分析/层次化注意力/流程工业智能优化Key words
collaborative optimization/graph-aware reinforcement learning/Laplacian spectral analysis/hierarchic-al attention/intelligent optimization of process industries引用本文复制引用
刘柢炬,王雅琳,刘晨亮,罗彪,桂卫华..基于多智能体强化学习的流程工业多操作参数协同优化[J].自动化学报,2026,52(1):78-90,13.基金项目
国家自然科学基金(U25A20466,92267205,62503507),湖南省自然科学基金(2025JJ60423,2025JJ10007),湖南省教育厅研究生教改项目(2025JGYB024)资助 Supported by National Natural Science Foundation of China(U25A20466,92267205,62503507),Natural Science Foundation of Hunan Province(2025JJ60423,2025JJ10007),and Hunan Pro-vincial Department of Education Graduate Education Reform Program(2025JGYB024) （U25A20466,92267205,62503507）
