Dynamic forecast method of logistics demand

Based on the JIT (Just In Time) management pattern, in order to improve the precision and the real time ability of logistics demand forecast, a compound algorithm was applied to design the dynamic forecast method of logistics demand. The compound algorithm that combines the Extended Kalman Filter with the Artificial Neural Network was used as the basic method for dynamic forecast. Dynamic influencing factors of logistics demand, such as sunlight, humidity,temperature and so on, were quantified and taken into consideration in the forecast process. With the assistance of Swarm simulation, we demonstrate that considering dynamic influencing factors in the forecast model will improve the precision and the real time ability of logistics demand forecast. Comparison of the forecast error between the general BP algorithm and the compound algorithm reveals the higher reliability of the proposed algorithm.