GA-BP Algorithm Application And Research In VMI Inventory Management

With increasingly fierce competition in the global market, the traditional business model has been difficult to meet market’s demand. The inventory management mode of Vendor Managed Inventory (VMI) has been widely used and rapidly developed under highly competitive environment. While in actually VMI inventory is often distributed in many places, companies need to store distributed inventory for centralized management. The main strategy of distributed VMI inventory management need companies to share real-time information with each, what’s more, accurate safety stock prediction can effectively reduce the consumption of inventory management. However, safety stock is affected by many uncertain factors, and there still has quite complex nonlinear relationship between the factors and decisions, while using neural networks to solve complex nonlinear problems have unique advantages compared with other methods. But traditional BP neural network also has some disadvantages, such as slow convergence, easy to fall into local optimum problem and so on, while GA-BP neural network has a more stable modeling model, more times to reach the preset target, less number of iterations to fit the data, more quickly to reach the preset goal, better data fitting effect.This paper studies the safety stock prediction method of VMI distributed inventory, analyses various prediction methods in detail, and selected GA-BP algorithm as the inventory forecasting method. Considering actual inventory management use distributed inventory mode frequently, while there is no need to make all stocks reaching safety stock point in distributed mode, allowing individual stock achieve zero inventory, which increases the uncertainty of the factors. In this paper, based on detailed analysis of the various uncertainties factors and through examples to prove that GA-BP algorithm has a good practicability. Using simulation technology to improve data fitting degree and improve prediction accuracy effectively. Finally, C#and Matlab mixed programming established VMI safety stock forecasting system, which implements GA-BP algorithm applied in VMI distributed inventory management, and fully demonstrated the uncertainty of factors in distributed inventory management model.