A multi-agent coordination of a supply chain ordering management with multiple members using Reinforcement Learning

Improving decision-making practices in a supply chain is a major source of competitive advantage in today´s uncertain business environments. There is strong evidence of success in the supply chain performance in cases with high coordination among echelons. The bullwhip effect is an important phenomenon in a supply chain, in which the order variability increases as orders move up in a supply chain. Reinforcement Learning (RL) is successfully applied to some dynamical and unpredictable domains. By surveying some efficient multi-agent RL models, this paper proposes a multi-agent coordination mechanism for a supply chain ordering management system with multiple members by the method of the RL. As the improvement to previous works using RL in the supply chain ordering management domain, the method proposed in this paper can be utilized to deal with multiple members in each echelon. As a result, the RL agent derives the maximal profit using the RL technique in the stochastic supply chain with multiple echelons and multiple members in each echelon.