摘要
为了解决城市交通流量预测问题,研究道路网中多断面同时作用的情况,先将各断面交通流时间序列在相空间中重构,以充分提取交通流中的相关信息,然后应用粗集理论的强定性分析能力对输入信息进行约简,消除了样本中的噪声和冗余。在此基础上,再利用支持向量机对约简信息进行预测。为了获得最优的预测精度,该方法还利用遗传算法对预测进行了优化。实例研究表明,该方法的预测效果令人满意,在交通控制领域具有较大的应用潜力。
In order to solve prediction problem of traffic volumes,this paper presented a scheme to forecast short-term traffic volumes of a road network. In this scheme,firstly,reconstructed the traffic volumes data of multi-road-cross-sections acquired from a road network in the phase space,and extracted correlative information in the traffic volumes richly,then reduced the input information by using the strong qualitative analysis ability of rough set theory,and removed the noise and redundancy in the samples. On the basis of it,using support vector machine,predicted reduced information. In order to obtain optimum predicting accuracy,used genetic algorithm to optimize prediction parameters. Practical case research shows that the predicted result of this method is famous,and this method has biggish applied potentials in the region of traffic control.
出处
《计算机应用研究》
CSCD
北大核心
2010年第10期3683-3685,3690,共4页
Application Research of Computers
基金
高等学校博士学科点专项科研基金资助项目(20060286005)
关键词
道路网
交通流量
相空间重构
粗集理论
支持向量机
预测模型
road network
traffic flow
phase reconstruction
rough set theory
support vector machine
prediction model
