基于Rough Set与灰色理论的公路货运量预测研究

发布时间：2018-09-19 12:52

【摘要】：世界经济发展已经进入全新时期,全球化经济正在日益向前推进,各经济体之间相互融合,相互影响。全球经济赖以生存的全球物流网络正在逐步建设,不断完善。我国正在组建快速、高效、全面的物流网络。目前我国已基本建设完成多形式、多渠道、全方位的交通运输物流网,这个运输物流网包括公路物流、铁路物流、水运物流、远洋物流和管道物流等多种物流运输方式,其中公路物流以其机动灵活、快速高效、运量大的优势成为物流网络中的主力军。国家在公路建设投资的控制和公路发展战略的制定中需要充分考虑公路带来的效益,公路效益来自公路运输,公路运输需要公路货运量的支撑,准确的预测结果对指导未来公路建设布局、未来经济发展有重要的意义。基于上述背景和目的,本文主要研究的问题是运用Rough Set理论和灰色理论相结合的方法分析、预测公路货运量。首先,本文在国内外研究现状的基础上,充分分析了公路货运量预测模型的特点,结合这些特点进行了相关理论知识的学习,为本文预测模型的选择奠定基础;其次,从宏观和微观两个方面对公路货运量的影响因素进行说明,建立较为完善的指标体系,运用灰色变权聚类和Rough Set理论进行影响因素的分析,根据获取规则预测公路货运量。具体来说,对指标进行数据统计,生成信息表,利用灰色变权聚类方法判别频率权,将信息表、公路货运量增长率、频率权结合生成决策表,利用Rough Set理论对决策表进行分析,提取规则,运用规则对公路货运量的增长率进行预测;接着,对传统灰色Verhulst模型进行改进,得到无偏灰色Verhulst模型。模型改进过程中运用到的思想为无偏GM(1,1)模型直接建模法,采取的方法为对原始序列作倒数生成,运用新生成的序列建立模型,并对传统灰色Verhulst模型和无偏灰色Verhulst模型自身误差进行分析;最后,通过实例验证说明模型改进方法的可行性以及改进后模型预测公路货运量的适用性。实例为兰州至中川2009-2015年公路货运量的模拟预测,分别运用GM(1,1)、传统灰色Verhulst模型、无偏灰色Verhulst模型三种模型进行对比分析。本论文的结论主要包括两方面:一方面,运用优势粗糙集理论对决策表进行约简、提取规则,预测公路货运量;另一方面,无偏灰色Verhulst模型消除自身固有的偏差,提高公路货运量的预测精度,说明改进后模型的可行性与适用性。
[Abstract]:The development of the world economy has entered a new period, and the global economy is advancing day by day. The global logistics network, on which the global economy depends, is gradually being built and perfected. China is building a fast, efficient and comprehensive logistics network. At present, our country has completed a multi-form, multi-channel, all-dimensional transportation logistics network, which includes road logistics, railway logistics, waterborne logistics, ocean logistics and pipeline logistics, and so on. Among them, highway logistics has become the main force in logistics network with its advantages of flexibility, speed and efficiency and large volume of transportation. In the control of highway construction investment and the formulation of highway development strategy, the state should take full account of the benefits brought by the highway. The highway benefit comes from the highway transportation, and the highway transportation needs the support of the highway freight volume. Accurate prediction results are of great significance to guide the layout of future highway construction and economic development in the future. Based on the above background and purpose, the main problem of this paper is to use the method of combining Rough Set theory and grey theory to forecast highway freight volume. First of all, on the basis of domestic and foreign research status, this paper fully analyzes the characteristics of highway freight volume forecasting model, combines these characteristics to study the relevant theoretical knowledge, and lays a foundation for the selection of this prediction model. This paper explains the influencing factors of highway freight volume from macro and micro aspects, establishes a relatively perfect index system, analyzes the influencing factors by using grey variable weight clustering and Rough Set theory, and predicts the highway freight volume according to the acquisition rules. Concretely speaking, the index data is counted, the information table is generated, the frequency weight is judged by the grey variable weight clustering method, the information table, the increase rate of highway freight volume and the frequency right are combined to generate the decision table, and the decision table is analyzed by using Rough Set theory. The rules are extracted to predict the growth rate of highway freight volume, and the traditional grey Verhulst model is improved to obtain the unbiased grey Verhulst model. The idea used in the process of model improvement is the direct modeling method of unbiased GM (1K1) model. The method adopted is the reciprocal generation of the original sequence and the establishment of the model by using the newly generated sequence. The error of the traditional grey Verhulst model and unbiased grey Verhulst model is analyzed. Finally, the feasibility of the improved model and the applicability of the improved model to forecast highway freight volume are verified by an example. An example is given for the simulation and prediction of highway freight volume from Lanzhou to Zhongchuan in 2009-2015. Three models, GM (1 / 1), traditional grey Verhulst model and unbiased grey Verhulst model, are used to carry out comparative analysis. The conclusion of this paper mainly includes two aspects: on the one hand, the advantage rough set theory is used to reduce the decision table, extract the rules, and predict the highway freight volume; on the other hand, the unbiased grey Verhulst model eliminates the inherent deviation. The feasibility and applicability of the improved model are illustrated by improving the forecasting accuracy of highway freight volume.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U492.313;N941.5

【参考文献】