Intelligent use of delayed information in the supply chain by artificial neural network

In global and large distributed systems, it is very important to exchange information correctly and efficiently. The supply chain is one such example. However, because of the high communication cost and the long network delay, information may be delayed and thus incorrect. Delayed information will affect the quality of decision making and the productivity of the system. An artificial neural network approach has been adopted in many forecasting applications and has proved to be very useful. We propose an artificial neural network approach for predicting the correct information from the delayed information in the supply chain. In this approach, a multilayer feedforward network trained by standard backpropagation is employed to realize the relationship between the delayed information and the correct information. To demonstrate the effectiveness of this approach, a simplified supply chain system is simulated and the effects of delayed demand information are examined. We compare our approach with the no-prediction approach and the perfect-prediction approach. Experiments show that our approach outperforms the no-prediction approach. Our approach exhibits comparable performance to the perfect-prediction approach when there is a small delay.