Research On Distributed Inventory Technology Of General Parts Products For Small Lots And Varieties

Because of the complexity of their management and the characteristics of their own products,small batch and multi-species general-purpose parts manufacturing enterprises are getting more and more attention.This paper proposes a recurrent neural network algorithm that can rely on long-term data relationships for inaccurate traditional inventory forecasting,and then proposes a coordinated allocation strategy for total inventory costs that are too high.This article mainly completed the following aspects:First of all,in order to solve the problem that the traditional time series forecasting method can not make full use of the long-term dependence between data,this paper studies a long short-term memory(LSTM)recurrent neural network algorithm for inventory forecasting research.The effectiveness of LSTM's inventory forecast is analyzed by simulation.On this basis,in order to solve the problem of overlapping input data and underutilization of the state of the neural unit,an improved algorithm based on LSTM is proposed for inventory prediction.The improved LSTM specifically improves the learning of complex data and makes full use of input data by adjusting the threshold structure.Through numerical simulation,the improved LSTM recurrent neural network has better effect on inventory prediction from multiple aspects such as images,charts and quality assessment methods.The simulation results show that compared with the original LSTM algorithm,the improved LSTM algorithm converges faster,the operation efficiency is higher,the model training time is shorter,the predicted value is closer to the true value,and its effect is higher than the other methods.Secondly,in view of the cost problem caused by the mutual transfer of goods between distributed inventory warehouses,this paper studies a single active transfer model.The central warehouse replenishes the regional secondary warehouse every other replenishment cycle.This model can alleviate the inventory backlog in some regional warehouses,but to a large extent,it cannot meet the needs of users.Other warehouses at the same level incur high inventory holding costs.In order to solve the problem of high inventory costs of the entire inventory system,this paper proposes a coordinated allocation strategy that can be replenished horizontally between secondary warehouses,which solves the problems of out-of-stock losses and reduces inventory holding costs.The model is established to coordinate the inventory level between each secondary warehouse,to minimize the overall cost within a certain period of time as the optimized objective function,and to minimize the allocation cost and increase through the reasonable inventory allocation method between the secondary warehouses.At the same time,it reduces the inventory cost and avoids the extra cost caused by blind scheduling.Finally,the model is solved to verify the effectiveness of the strategy.Finally,on the basis of the above research,this paper develops a distributed inventory system based on SpringBoot and other Java enterprise frameworks to guide the production of the general parts manufacturing enterprise.The system contains a total of six major subsystems.The detailed design,code implementation and complete testing of the system functions are carried out.