| Port logistics can be regarded as an indispensable part of the development of modern logistics.It is an engine of economic development in a city and a radiating region,and an inevitable product of economic development in the times.China’s port logistics is in a period of rapid development,and studying port logistics system planning has become an important topic.This article summarizes the literature on forecasting port logistics demand at home and abroad,and summarizes that many researchers usually choose the impact indicators commonly used in the industry when forecasting port logistics demand;and when choosing the method of logistics demand forecasting,focus on the pursuit of port logistics forecasting accuracy and ignore the essential difference of each demand forecasting method.Based on the above problems,this article takes Qingdao Port as the research object and conducts logistics demand forecasting research with a view to obtaining more accurate and time-sensitive forecasting results.Main tasks as follows:(1)The factors affecting port logistics demand are analyzed.After the field investigation,combined with the current development status of Qingdao Port Logistics,through the index discriminant analysis method and the Pearson correlation coefficient method,the influencing factors of port logistics demand that can reflect the research object and have a high degree of correlation are selected.And establish the Qingdao Port logistics demand forecast index system to lay the foundation for Qingdao Port logistics demand forecast research.(2)Port logistics demand forecast model is established.By comparing the advantages and disadvantages of various prediction methods,it is found that the grey prediction model is simple to calculate and requires less data;BP neural network is suitable for systems that deal with complex internal problems and do not affect the overall prediction effect when local neurons fail.Therefore,in order to solve the problem of poor data and non-linearity of the logistics demand forecast of Qingdao Port,this paper assigns specific weights to the two forecasting models of grey forecasting and BP neural network by least squares,and builds a combined prediction model of GM(1,1)-BP neural network.(3)After conducting a field survey of Qingdao Port,the article selects the relevant data of the port logistics demand influencing factors from 2000 to 2014 to train the prediction model.The data from 2015 to 2019 verifies the rationality of the model.The relative error of the predicted value of the combined prediction model of GM(1,1)-BP neural network is 3.25%.In the same kind of prediction environment,use other models to reproduce.Through comparison,it is found that the GM(1,1)-BP neural network prediction model selected in this paper has better prediction results.(4)The establishment of the Shandong Port Group and the new establishment of the Shandong Pilot Free Trade Zone have increased the opening of Qingdao Port to the outside world.However,the global COVID-19 outbreak has caused a decline in consumer demand and even affected the normal operation of the entire international trade market,Which also has a certain impact on the throughput of Qingdao Port.In order to reduce the impact of these external factors on the forecast results,the article uses the Markov forecast model to convert the forecast value of the logistics demand of Qingdao Port into the corresponding forecast interval,so that the forecast results are more scientific and credible.This article aims to solve the problems of current port logistics system planning,and builds a reasonable forecast model to forecast the logistics needs of Qingdao Port.The experimental results show that the GM(1,1)-BP neural network combined forecasting model selected in this paper has better applicability in the logistics demand forecast of Qingdao Port.It not only provides new ideas for port logistics demand forecasting,but also provides data support for the rational allocation of port logistics resources. |
PDF Full Text Request