Supply Chain Economic Decision-Making Based on Reinforcement Learning Algorithm

Supply chain management is rife with challenges such as demand volatility, poor inventory control, inefficient resource allocation, etc. These factors have a huge impact on the economic decisions of companies. This paper focuses on addressing these issues in order to optimize supply chain economic decisions through reinforcement learning algorithms. The research method is to establish a decision-making model based on reinforcement learning, design the state space, action space and reward mechanism. After that, a simulation of the real supply chain scenario is built in the simulation and trained using Q-learning and deep Q-network (DQN) algorithms to optimize inventory levels and order strategies. Experimental results show that after 1,000 iterations of the reinforcement learning model, the inventory turnover rate increased from the original 4.5 to 7.8, and the average cycle cost was reduced by about 15%. This improves the overall efficiency of the supply chain. The conclusion shows that reinforcement learning algorithms provide a feasible solution for the supply chain economic decision-making support process, which can better allocate resources and manage risks in a complex and changing market environment.