| In recent years,with the vigorous development of the logistics industry,more and more research has been conducted on the logistics park,which is the core of the logistics industry.However,the current research is more concentrated on the location layout and system construction of the logistics park,and there is less research on improving the microcirculation inside and outside the logistics park to improve the efficiency of freight flow.For the logistics park,the effective way to improve the flow efficiency is to predict congestion,and then break the bottleneck of the park operation and improve efficiency.The current congestion prediction technology is rarely used in logistics park scenes,and it is difficult to directly transplant other scene prediction methods for congestion prediction.At the same time,on the basis of solving the large amount of traffic data,the business attribute requirements of congestion prediction in logistics parks must also be considered.The Apache Spark platform is currently an option to achieve congestion prediction and has advantages that traditional frameworks cannot match.In order to realize the congestion prediction of the logistics park and improve the circulation efficiency of the park,this paper proposes a congestion prediction system of the logistics park based on the Apache Spark framework.Aiming at the traffic characteristics of the logistics park,a logistics park monitoring network was proposed in the congestion prediction system to better collect traffic data for the micro-circulation in the logistics park and the surrounding small area.Considering the business characteristics of the logistics park,a logistics park related Congestion factor.Asynchronous optimization is performed on the logistic regression algorithm on the predictive modeling method,so that it has stronger performance in big data scenarios.In order to realize the congestion prediction of the logistics park and improve the circulation efficiency of the park,this paper proposes a congestion prediction system of the logistics park based on the Apache Spark framework.The current common scenarios for congestion prediction are cities and highways,and there are few studies applied to logistics parks.In order to improve the accuracy of the logistics park congestion prediction and its adaptability to the logistics park business,this paper proposes a congestion prediction model based on the logistics park monitoring network.Through better traffic data collection for the microcirculation in the logistics park and surrounding small areas,design the congestion coefficient associated with the business characteristics of the logistics park,improve the accuracy and pertinence of prediction,help solve the congestion in the logistics park,and improve cycle.In order to better handle the massive real-time traffic data in the logistics park congestion prediction model and improve the prediction speed,this paper proposes the use of the Spark framework to implement the congestion prediction method based on asynchronous logistic regression applied to the logistics park congestion prediction model.By distributing data to multiple nodes and calculating at the same time,the operation speed of the logistic regression algorithm is improved,so that it has stronger performance in the case of big data.Finally,this paper designs and implements a logistics park congestion prediction system based on Spark.The system is divided into data acquisition layer,data processing layer,model storage layer and data visualization.Use Spark to build big data processing services,My SQL to build relational database storage services,Flume,Kafka as data transmission middleware,complete system Web services based on Spring Boot framework.The function of the system has been tested in actual application scenarios,which proves that the congestion prediction system has strong practicability. |
