A new adaptive inventory control method for supply chains with non-stationary demand

Inventory control plays an important role in production and supply chains management. As the cycle of productions becomes shorter and customer demand becomes more unstable than before, the ordinary inventory control method can not deal well with the problem. Adaptive inventory control method receives attention of both academicians and practitioners. But none analytical solutions to safety stock is given when the stochastic demand distribution is unknown. In this paper, using the reinforcement learning technique and the BP neural network, we propose a new adaptive inventory control method for supply chains consisting of one supplier and multiple retailers. In our approach, stochastic demand is non-stationary distributed and distributions function is not required. Simulation results show that the method we propose outperforms the method in related literature, thus it provides a possible for inventory control in situations with highly demand fluctuation.