Simulation and Optimization for Supply Chain Based on Multi-agent Reinforcement Learning: A Case Study on a Large-scale Refinery

摘要

A case study on a large-scale refinery is implemented in this paper.Considering the plenty of stochastic factors existing in real life,general-purpose simulation platform ARENA is employed to model the complex supply chain of this refinery and obtain the system's performance indices.With VBA and Object Oriented Programming technology,a kind of architecture is proposed to integrate simulation with optimization.Then a Multi-agent Reinforcement Learning algorithm is designed to optimize the ordering and distributing policies of the refinery.Research results show that the methodology proposed can effectively solve the optimization problems existing in real-life and complicated supply chain.