| In recent years,the research on time series has a huge impact on the development of time series.With the continuous development of global automobile industry,for the automobile enterprises,the prediction of the output of various models has attracted more and more attention.Accurate production forecast can make enterprises plan strategic reasonably,reduce production cost,avoid production waste,and ensure the steady development of each enterprise in the supply chain.LSSVM(least square support vector machine)algorithm is an optimized algorithm based on SVM(support vector machine).It has the advantages of simple model,high learning efficiency and low training cost.This feature is suitable for the production prediction of vehicle models,so this thesis uses this algorithm to study.First of all,it analyzes the method of making production plan of Plant D from the aspects of business process,production,logistics,order and so on.Then,the basic principle of LS-SVM algorithm is studied by analyzing a variety of time series models.In this thesis,the historical data of each vehicle type in Plant D are collected,and different strategies are selected for different types of samples to reconstruct the vector space.The LSSVM algorithm based on genetic algorithm optimization parameters is used to predict the time series of each group of data,and the prediction results are verified.According to the prediction results of time series with different characteristics,the corresponding conclusions are obtained,and the optimization scheme of production planning related business is further given,including long-term production planning,medium and short-term production planning,order sequencing,production department scheduling,supply chain related business,etc.From the experimental results of the method selected in this thesis,compared with the traditional production forecasting method,the LSSVM time series prediction method based on SaGA has the advantages of higher prediction accuracy,faster prediction speed and more simplified process.According to the experimental results,this thesis gives the optimization suggestions of the production planning related fields.A large number of experiments show that the LSSVM time series prediction model based on SaGA can be further promoted in the vehicle industry. |
