Research On Matching Strategy And Path Optimization Of Network Freight Platform Based On Spatial Feature Analysis

The information aggregation and resource integration capabilities of the Internet make the platform transformation an important direction of the transformation and upgrading of transportation industry.Although the new format of online freight platform has entered the field of transportation services and taken shape,the complexity and heterogeneity of transportation organization and the efficiency differences of transportation subjects have largely restricted the smooth operation of online freight platform transactions.Therefore,from the two dimensions of transportation mode constraints of transportation capacity and effective collection of transportation capacity,vehicle-cargo matching and path selection can be optimized more precisely to reduce cost and increase efficiency,and improve the competitiveness of the online freight platform.Taking the online freight platform as the research object,this dissertation focuses on the theme of matching strategy and path optimization of network freight platform based on spatial feature analysis".Firstly,in order to identify the operational characteristics of the online freight platform and the spatial characteristics of the transport behavior,a framework of mobile pattern analysis is established to mine the spatial pattern characteristics of the transport vehicles in the online freight platform through the method of network base sequence,while using the Kmeans++ clustering algorithm to classify the travel chain behavior patterns of the transport vehicles in the platform and to identify potential vehicle organization patterns through the correlation analysis method of spatial activities and behavior patterns;Secondly,the performance screening of potential transportation vehicles is realized from two perspectives of operational efficiency and safety by constructing a user behavior identification model of transportation vehicles of online freight platform.By establishing the evaluation index system of the behavior recognition model,the K-means++algorithm is used to cluster the indexes of the two dimensions of operation efficiency and safety respectively,after which the mean value is supplemented for reasonable screening,and the utility value of the operation vehicles is calculated through the entropy weight method of comprehensive evaluation index weights for secondary selection as the reference basis of the final effective capacity collection;Thirdly,under the constraints of vehicle operation mode and the selected effective capacity set,based on the cost optimization strategy,a multi-way hybrid loading and unloading model for vehicle cargo matching of online freight platform is constructed.The model takes the minimum total vehicle driving distance and the minimum vehicle utilization scale as the dual objective constraint function,and takes into account the types of vehicles and goods,vehicle path,maximum vehicle carrying weight,vehicle operation time and time window as the constraint factors.Then,the particle swarm optimization algorithm is used to solve the model by using the real number coding method to generate particles and modify them,and the particle swarm optimization algorithm(PSO)is used to solve the continuity problem of multi-path hybrid loading and unloading;Finally,for the multimodal transport online freight platform,considering the complexity of the multimodal transport network,the K-shell-based multimodal path optimization algorithm is constructed in combination with the super network theory,and an empirical analysis is conducted with the multimodal network in the Yangtze River Delta region as an example to verify the feasibility of the network model and the path algorithm solution in solving the practical problems of large-scale network.The main innovations of this dissertation can be summarized as follows:(1)A new method for the identification and selection of transportation capacity is proposed.On the one hand,the inherent transportation organization mode is extracted based on the vehicle operation characteristics of the transportation subject as the vehicle constraint in the vehicle cargo matching decision-making;On the other hand,based on the two dimensions of transportation efficiency and transportation safety,the method of clustering and entropy weight is used to evaluate and select transportation entities,which provides an effective transportation capacity set for vehicle and cargo matching decision-making;(2)The optimization model of multi-way mixed loading and unloading with the minimization of vehicle cargo matching cost of network freight platform is constructed,and the effective optimization of vehicle selection,transport paths between collection and delivery points is realized with full consideration of the basic information of vehicles and transport organization mode,as well as the running time and time windows and other constraints;(3)A multimodal transport super-network model is constructed,and a multimodal transport optimal path generation algorithm based on improved K-shell algorithm is designed,which solves the problem of multimodal transport network freight platform path optimization.