Research On Demand Forecast Of Molybdenum Mineral Logistics In China Based On Improved LSSVM

With the rapid development of China's metallurgical industry,chemical industry,agriculture and aerospace.As an important alloy material,the demand for molybdenum is increasing,and the flow rate of molybdenum fine minerals is increasing rapidly.In addition,due to the large scale,wide area and long distance of molybdenum resource products in China,the demand for logistics services is even greater.The construction of the original molybdenum fine mineral flow service system failed to fully consider the future demand of molybdenum fine mineral flow.Therefore,accurately grasping the spatial geologic structure of molybdenum concentrate demand and predicting the total demand of logistics is a reasonable plan for molybdenum fine mineral flow system.To achieve an important guarantee for the balance of supply and demand of molybdenum fine mineral resources.At present,the solution to the problem is mostly machine learning,and LSSVM is more suitable for the prediction of small sample complex systems with logistics requirements.However,the prediction accuracy of LSSVM depends on the setting of kernel parameters and penalty parameters.Therefore,this paper introduces DACPSO,CAS,DE-BA algorithm for LSSVM parameter optimization,and establishes molybdenum mineral flow demand prediction model to realize molybdenum in China for the next three years.This article has mainly completed the following work:(1)The status quo of molybdenum fine mineral flow in China and the knowledge system related to molybdenum fine mineral flow demand are analyzed in detail.Including: molybdenum fine mineral flow and logistics demand concept and characteristics analysis,molybdenum fine mineral flow node function,spatial structure and transportation pattern analysis.The basic steps of molybdenum mineralstream demand measurement indicators and logistics demand forecasting were determined.At the same time,collect the logistics statistics related to the freight volume and import and export volume of each logistics node,and lay a solid theoretical foundation and data foundation for the demand forecast of molybdenum mineral flow.(2)From the macroscopic perspectives of industry chain,economy and society,the main factors affecting the demand of molybdenum mineral stream in China are analyzed,and the relevant data of main influencing factors are collected.On this basis,the construction of molybdenum mineral stream demand forecasting index system is completed.When dealing with the problem of index selection,the gray correlation method is used to analyze the relationship between the relevant influencing factors and the molybdenum mineral stream demand,and the importance degree of each influencing factor is obtained.Then,the main influencing factors are eliminated by nuclear principal component analysis.The multiple correlations provide the necessary foundation for the establishment of the next logistics demand forecasting model.(3)A model for predicting molybdenum mineral flow demand based on improved LSSVM was constructed.Aiming at the complexity,nonlinearity and randomness of the molybdenum mineral flow system,the LSSVM model is used to solve the logistics demand prediction problem.On this basis,for the LSSVM parameter optimization time-consuming and poor effect,the DCSVM,CAS and DE-BA algorithms are used to optimize the LSSVM kernel parameters and penalty parameters,based on DACPSO-LSSVM,CAS-LSSVM,DE-BA-LSSVM molybdenum mineral flow demand prediction model.(4)An empirical study on the demand forecast of molybdenum fine minerals in China.A total of 28 sets of data from 1991 to 2018 were selected to predict the demand for molybdenum concentrate flow in the next three years.The data samples are divided into training samples and test samples and the data is preprocessed.The RBF function with strong generalization ability is selected as the kernel function,the mean square error is defined as the fitness function,the parameter settings are adjusted,and the LSSVM basic model and optimization are performed.The model establishment process,and the corresponding training results and test results areobtained and compared.The results show that the prediction effect of DE-BA-LSSVM model is obviously superior.Therefore,the DE-BA algorithm is finally determined as the LSSVM parameter optimization method,and the DE-BA-LSSVM model is used as the molybdenum mineral flow demand prediction model.The modeling time is as low as 3s,and the prediction accuracy of the built model can reach 98.42%.In addition,the extrapolation application of the model indicates that the demand for molybdenum fine minerals in China is on the rise in the next few years.Based on the analysis of the current status of molybdenum mineral stream transportation,based on the forecast of molybdenum concentrate freight volume,LSSVM,group intelligent optimization algorithm and real-time logistics demand forecasting problem are combined to scientifically and reasonably propose LSSVM parameter optimization method and establish A molybdenum mineral flow demand prediction model.The prediction results are used as the basis for the planning of molybdenum fine mineral flow system.It provides a reasonable decision-making basis for the government to grasp the demand intensity of molybdenum fine mineral flow,ensure the relative balance of supply and demand of molybdenum fine mineral flow service and the future research of molybdenum concentrate strategic reserve.