Research On Qingdao Logistics Demand Forecast Based On ACO-SVM

As an important part of the national economy and the most economical service model in the process of industrialization,the logistics industry has developed rapidly on a global scale.In recent years,with the rapid development of urban economy,the development of modern logistics,optimization of economic structure,improvement of investment environment,and improvement The overall competitiveness of urban economy has important strategic significance.The study of urban logistics planning has become an important issue,and successful planning and decision-making depend on the accuracy of prediction.Therefore,scientific and reasonable prediction is the key link of successful logistics planning,in other words.The primary problem in logistics planning for urban logistics is the forecast of logistics demand.This paper summarizes the development of logistics demand forecasting research at home and abroad,analyzes the connotation and characteristics of logistics demand,and studies the urban logistics demand forecast with Qingdao as the background.Mainly done the following aspects:(1)The factors affecting urban logistics demand are analyzed.The correlation analysis between freight volume and urban economy is illustrated by SPSS correlation analysis.Combined with the development status of Qingdao logistics,the main factors affecting Qingdao logistics demand are studied,and the logistics demand forecasting index system of Qingdao is constructed..Through the entropy weight method and grey correlation analysis,the indicators are quantitatively analyzed,and the index system is reduced.The influencing factors with high correlation with Qingdao logistics demand are selected for the logistics demand forecasting research in Qingdao,which lays a foundation for logistics demand research.(2)The logistics system is a complex nonlinear system.For the characteristics of few data samples,lack of internal connection and regularity,by comparing the advantages and disadvantages of various prediction methods,combined with Support Vector Machines(SVM)to solve limited samples,The unique advantages of nonlinear functions and multi-dimensional pattern recognition are the application of support vector machines to urban logistics demand forecasting.The experimental results of classification accuracy of four different kernel functions are carried out.In this paper,the RBF kernel function with the highest classification accuracy is selected.On this basis,Ant Colony Optimization(ACO)is designed to affect the prediction accuracy of support vector machine.The g-parameters were optimized.The experimental results showed that the accuracy was 97.28%.The optimized support vector machine prediction model was established,which provided an effective new method for Qingdao logistics demand forecasting.(3)Based on the index data of Qingdao City from 1999 to 2017,the empirical research on logistics demand forecasting is carried out for the training set.The prediction results show that the ant colony algorithm based optimization support vector machine prediction model is obviously better than the RBF neural network model in predicting accuracy,thus verifying The high adaptability and fit of the model in this paper.This paper makes reasonable forecast of logistics demand,helps to grasp the structure of urban logistics demand,and effectively promotes the balance between urban logistics service demand and supply.It is of great significance to carry out high-quality urban logistics planning and improve the efficiency of urban logistics operation.Qingdao is striving to build a new pattern of kinetic energy conversion and development.It is estimated that by 2022,the logistics demand will reach more than 3 million tons.Through theoretical and quantitative analysis to predict the scale of logistics demand in Qingdao,it can effectively provide methods and experience for the future research of Qingdao logistics industry.