Abstract:

Bullwhip effect refers to the phenomenon of demand variation and gradual amplification in the supply chain， which will bring high costs to supply chain enterprises. Therefore， alleviating the bullwhip effect has always been one of the core tasks of supply chain management， but the existing methods still cannot completely eliminate bullwhip effect. Introducing artificial intelligence methods into the supply chain to create a smart supply chain is one of the main methods for supply chain management in the information age， and it is also the main source of ideas to study the bullwhip effect. To explore new ways to alleviate the bullwhip effect， a four-level supply chain consisting of a retailer， a wholesaler， a distributor and a manufacturer is constructed， the deep reinforcement learning technology in artificial intelligence is introduced into supply chain management， and a human-machine collaborative intelligent decision-making robot is designed， used to improve the decision-making of distributors in the supply chain. The results of the smart experiment show that the indicators such as “variance ratio， average cost and service level” of each node in the supply chain have been comprehensively improved， which means that the bullwhip effect has been greatly alleviated. A new perspective and approach for studying the bullwhip effect in supply chain management is provided， and supply chain management is endowed with the value of intelligent decision-making.

Key words: bullwhip effect, human-machine collaborative, intelligent decision-making robot, supply chain management