Asynchronous Multi-Agent Collaborative Framework for Integrated Production and Procurement Optimization in Refinery

The refinery industry operates as a highly complex system characterized by dynamic interactions among numerous processes, resources, and decision-making strategies. Within this context, production planning and crude oil procurement management are critical to ensuring operational efficiency and profitability. However, traditional optimization approaches often neglect the distinct decision-making time scales of these two functions, limiting their ability to address market dynamics and supply chain complexities effectively. To overcome these challenges, this study introduces an asynchronous multi-agent collaborative optimization framework that enhances the coordination between production planning and crude oil procurement in refinery operations. By enabling production and procurement agents to operate on independent time scales, the framework adapts to fluctuations in crude oil prices and variations in product demand. The study further extends the classical multi-agent reinforcement learning algorithm into an asynchronous paradigm, introducing Async-MAPPO, which allows agents to independently optimize decisions using real-time data. This approach mitigates operational delays and enhances adaptability. Experimental evaluations demonstrate that the proposed method significantly improves production efficiency, reduces operational costs, and strengthens the refinery’s resilience to market fluctuations, underscoring the critical role of asynchronous optimization in refinery operations. Note to Practitioners—Refinery operations involve the intricate coordination of production planning and crude oil procurement, both of which operate on distinct decision-making time scales. This disparity in time scales often complicates efforts to achieve seamless coordination, especially in the face of uncertain disturbances such as fluctuating crude oil prices, variable product demands, and supply chain complexities. This paper proposes an advanced multi-agent reinforcement learning (MARL) framework coupled with an asynchronous collaborative decision-making mechanism. By enabling production and procurement agents to make decisions independently yet collaboratively across different time scales, the proposed framework addresses the dynamic nature of refinery operations. The asynchronous collaboration mechanism allows each agent to respond promptly to real-time data, reducing delays and enhancing adaptability to market fluctuations. This approach provides practitioners with a novel tool for managing the complexities of refinery operations in the future era.