Design And Implementation Of Intelligent Auto Parts Purchase-Sales-Inventory System

With the rapid development of China’s economy,the auto market has also grown rapidly.As the cornerstone of the auto industry chain,the auto parts market has also ushered in unprecedented prosperity.Due to the complexity auto parts species,auto parts dealers mostly use the purchase-sales-inventory system to manage daily business data,at the same time,current purchase sale and stock management system can only carry out simple records and statistics on business data,and they are lacking of business data analysis,forecasting function,so auto parts dealers mainly based on experience to manually predict parts sales,and thus make procurement plans.However,with the growing of business scale,the error of manual forecasting of parts sales is getting bigger and bigger,which makes it impossible for merchants to make reasonable purchasing plans,which leads to inventory backlog or inventory shortage,and ultimately causes economic losses for merchants.The purpose of this thesis is building an accurate and efficient parts sales forecast model,and embed it into the purchase-sales-inventory system used by merchants daily,to help them make reasonable business decisions,so as to create higher economic benefits.Firstly,the thesis investigates the research both domestic and foreign on purchasesales-inventory systems and sales forecasts,and we clarify the primary problems that the sales forecasting module of this thesis needs to solve.On this basis,the requirement analysis and UML use case model is accomplished.The main functional modules of the system include account management,purchase management,sales management,inventory management and sales forecast to meet the requirement of all links of business activities.Secondly,according to the requirements analysis,each functional module is designed.In the sales prediction module,a prediction model based on iba-lstm is constructed: Aiming at the problem of bat algorithm that the overall quality of the bat population is low due to the random initialization of bats,we designs an improved bat algorithm(IBA),which only accepts individuals whose fitness value is higher than the average value of the population when initializing the population,so as to ensure the quality of the initial population;At the same time,IBA is used to optimize the initial weight and bias of LSTM model,and the optimal weight and bias are used to construct and train LSTM network to improve the performance of the model.The experimental results show that the model can provide better prediction results than LSTM model and ba-lstm model.Finally,using layui,spring boot and tensorflow framework,each functional module is implemented,and the function and performance of the system are tested.The test results show that each module can run normally,and the response speed meets the actual requirements,and the system achieves the expected design goal.