基于强化学习的牛鞭效应对策模型

Solution model for the bullwhip effect in supply chain based on reinforcement learning

在不确定供应链环境中提高供应链节点决策能力是提高竞争优势的主要方式。需求波动沿供应链从下游到上游逐级放大的牛鞭效应是供应链管理的重要问题。在综合分析供应链牛鞭效应的成因以及现有对策基础上,针对多层、多节点供应链模型,分析并运用强化学习理论和方法,提出有助于减弱牛鞭效应的强化学习算法。该强化学习算法能够用在不确定、多层、多成员供应链环境中,可取得整体最大效益,从而提高供应链的运作效率。

In stochastic supply chain environments,improving decision-making practices is a major approach to improve competitive advantage. The bullwhip effect is a phenomenon in a supply chain, in which the order variability increases as moving upstream in the supply chain. By surveying major causes to the bullwhip effect and its previous solutions, a reinforcement learning method is proposed for minimizing the bullwhip effect in the supply chain model with multiple echelons and multiple members in each echelon. As an improvement to previous works, the proposed reinforcement learning method can be utilized to deal with multiple members in each echelon and derive the maximal profit in the stochastic supply chain.