基于灰色残差GM（1，1）模型的道路交通量预测的研究

道路交通体系是一个多因素、多层次、多目标的复杂系统。其中交通量信息系统具有明显的层次复杂性，结构关系的模糊性，动态变化的随机性，指标数据的不完全和不确定性。由于技术方法、人为因素、自然环境变化的影响，造成各种数据误差、短缺甚至虚假现象，系统的作用机制不明确，系统的状态、结构、边界关系难以精确描述，属于典型的灰色系统。在作量化、模型化、实体化研究时，能作为反映系统主要动态特征的数据是很少的。由于环境对系统的干扰，系统信息中原始数据序列往往呈现离乱情况，离乱数列即为灰色数列或称灰色过程，灰色理论利用那些较少的或不确切的表示系统行为特征的原始数据序列作生成变换后建立微分方程，对灰色过程建立的模型称为灰色模型（Greymodel），简称GM模型。本文从理论上介绍了GM（1，1）模型和灰色残差GM（1，1）模型建立的一般过程，然后将其应用于交通量预测的实际例子中。预测结果表明，该方法是可行的。

Research of Traffic Volume Prediction Based on GM(1,1)Model

Road transportation system is amulti-factor, multi-level andmulti-target complex system. Traffic information system obviously has a feature of layer complexity, ambiguity in the relationship among structures, dynamic random change, and incompleteness and uncertainty of the indicator data. Due to the technical methods, human factors, the impact of the natural environment change, all the kinds of data error, or even false shortage phenomenon, the system mechanism is not clear, and the system state, structure, and the border relation were difficult to accurately be described. So, it is a typical grey system. In the research, the data reflecting the system＇s major dynamic characteristics is few. Since the interference of the system environment, the raw data of the information system often appear discrete. Discrete series or called grey series process, which is used to express the system behaviour, was adopted to form differential equations. The model established by the grey process known as the grey model. This paper introduces the general forming course of GM（1,1）model, then apply it to traffic volume prediction. The result indicates that the method is right.