基于GA-BP算法的公路货运定价模型研究

发布时间：2018-03-30 05:17

本文选题：公路货运　切入点：运价制定　出处：《浙江工商大学》2017年硕士论文

【摘要】：公路货运是指在公路上,主要运输工具为货车的一种运输方式,是交通运输系统的重要组成部分。截止到2015年,我国高速公路总里程数已达457.73万公里,居世界第一,仅2015年公路货运量达315亿吨,占总货运量的76.8%。"无车承运人"是公路货运的一种新型商业模式,是指无自有车辆的公司或单位,进行承接物流运输项目,随着"互联网+物流"模式的逐渐推广,大量的传统货运企业向"无车承运人"企业转型,为了提高新型货运企业与实际承运商之间合作效率,在双方协定运价时,提供一个合理的参考运价具有重大意义。本文以公路货运的运价制定问题为研究对象,使用多元线性回归模型、BP神经网络模型和基于遗传算法优化BP神经网络模型(简称为GA-BP模型)进行研究。主要研究工作及成果总结如下:(1)综述了公路货运行业现状,提出了公路货运运价制定问题,及其研究的价值与意义,综述了该问题研究现状;介绍了相关启发式算法,并进行比较。(2)介绍了 "无车承运人"概念的由来,及其运作模式;针对公路货运运价制定问题,结合对无车承运人企业的调研,对影响运价的因素进行分析选取,并进行灰色关联度分析,最终选取了国家经济发展水平、燃油价格、市场需求量、货物重量、路程,并在此基础上建立了多元线性回归模型。(3)由于BP神经网络对非线性关系的有很强的适用性,因此将BP神经网络算法应用于公路货运运价制定问题,增加考虑非线性影响因素车型和承运商,构建7-15-1型的神经网络结构。(4)针对BP神经网络的收敛速度慢缺点,本文使用LM算法优化;针对BP神经网络算法的易陷入局部最优解的缺点,本文将遗传算法与BP网络算法结合,建立基于GA-BP算法的运价制定模型,采用遗传算法优化BP网络的初始权值与阈值,极大提高了算法搜索全局最优解的能力。(5)研究了公路货运模型的应用。根据M货运公司提供的实际经营数据,分别对多元线性回归模型、BP神经网络模型、GA-BP模型进行应用研究。通过实例分析验证了本文提出的基于GA-BP算法运价制定模型的有效性与实用价值。
[Abstract]:Highway freight is a kind of transportation mode in highway, where the main means of transport is a freight car. It is an important part of the transportation system. By 2015, the total mileage of expressway in China has reached 4.5773 million kilometers, ranking first in the world. In 2015 alone, the volume of goods transported by road amounted to 31.5 billion tons, accounting for 76.88 percent of the total cargo volume. "Carrier without vehicles" is a new business model for road freight transport, which refers to companies or units that do not have their own vehicles to undertake logistics and transportation projects. With the gradual promotion of the mode of "Internet logistics", a large number of traditional freight enterprises have been transformed to "carless carriers". In order to improve the efficiency of cooperation between new freight enterprises and actual carriers, when the two parties agree on freight rates, It is of great significance to provide a reasonable reference rate. Using multiple linear regression model and BP neural network model based on genetic algorithm optimization BP neural network model (referred to as GA-BP model). The main research work and results are summarized as follows: 1) the current situation of road freight industry is summarized. This paper puts forward the problem of highway freight freight pricing, and the value and significance of the research, summarizes the current research situation of the problem, introduces the relevant heuristic algorithms, and compares the concept of "car-free carrier", and introduces the origin and operation mode of the concept of "car-free carrier". In view of the problem of road freight freight tariff formulation, combined with the investigation of non-vehicle carrier enterprises, the factors affecting freight rate are analyzed and selected, and the grey correlation degree analysis is carried out. Finally, the national economic development level and fuel price are selected. Market demand, cargo weight, distance, and on this basis, a multivariate linear regression model is established. Therefore, the BP neural network algorithm is applied to the pricing problem of road freight transportation, and considering the nonlinear influence factors of vehicle type and carrier, the neural network structure of 7-15-1 type is constructed, which aims at the slow convergence speed of BP neural network. In this paper, LM algorithm is used to optimize, and the BP neural network algorithm is easy to fall into the local optimal solution. In this paper, the genetic algorithm and BP network algorithm are combined to establish the pricing model based on GA-BP algorithm. Genetic algorithm is used to optimize the initial weight and threshold of BP network, which greatly improves the ability of searching global optimal solution.) the application of road freight model is studied. According to the actual operating data provided by M freight company, The application of GA-BP model, a multivariate linear regression model, is studied, and the validity and practical value of the model based on GA-BP algorithm are verified by an example.
【学位授予单位】：浙江工商大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F542.5;TP183

【参考文献】