The Study Of Road Freight Rate Model Base On GA-BP Algorithm

Road Freight,as a part of transportation system,is particularly for transporting goods on the road by trucks.Until 2015,China’s total highway mileage has reached 4.57 million kilometers,ranking first in the world,and the cargo volume of road freight has reached 31.5 billion tons,accounting for 76.8%of total cargo volume.Car-free carrier is a new road freight business model,refers to the company without its own vehicles,to undertake logistics projects.With the mechanism of logistics with Internet becoming more mature,a increasing number of traditional freight companies tend to car-free carrier.In order to improve the cooperation efficiency between the new freight carriers and the actual carriers,it is of great significance to provide a reasonable reference freight rate when negotiating.In this paper,it regards the road freight rate as research object,and uses BP neural network model,genetic algorithm based BP neural network model(GA-BP model)to solve relevant problems.The main research results are summarized as follows:Firstly,it summarizes the current situation of road freight transportation industry,puts forward road freight rate problem,analyzes its situation and research significance,and introduces relevant heuristic algorithms.Secondly,it introduces the origin of Car-Free Carrier and its operation mode.After investigating several Car-Free Carrier companies for road freight rate problem,filtering acting factors and analyzing with the method of gray correlation,it finally takes the level of national economic development,fuel prices,market demand,cargo weight and distance as its factors for the model.Base on this,it builds a multiple linear regression model.Thirdly,Since the effective applicability of BP neural network on nonlinear relationship,BP neural network algorithm is applied to the problem,considering truck type and carrier.Then a neural network structure type of 7-15-1 is constructed.Fourthly,it uses the LM algorithm to solve slow-convergence problem for BP neural network algorithm.Combining genetic algorithm and BP neural network algorithm,it constructs GA-BP algorithm to solve falling into local optimal solution problem.It proved that optimizing the initial weights and thresholds of the BP network using genetic algorithm greatly improves the ability of the algorithm to search the global optimal solution.Lastly,it studies the actual application of road freight rate model.According to data from M freight company,the comparison among multivariate linear regression model,BP neural network model and GA-BP model turns out that the proposed road freight rate model based on GA-BP algorithm are valuable and effective.