摘 要进入 21 世纪以来，人们的社会生活水平逐渐提高，随着车辆在大多数城市居民中的普及，城市交通拥堵问题恶化迅速。为切实地解决这个问题，合理地安排交通运行，提高交通设施的使用效率，必须进行交通流的合理分配以及信号灯的优化配置，准确的短时交通流量预测是解决这些问题的重要依据。近年来 BP 神经网络被广泛应用于预测研究。本文设计了三层 BP 神经网络用于短时交通流量预测。将预测日前三天同时段的交通量作为输入信号，预测当日某时段的交通量。文中用江门市区小时交通量数据进行了仿真实验，结果表明该用方法进行短时交通流量预测是可行的。最后，对预测结果的准确性进行分析和讨论。关键词 短时交通流量；BP 神经网络；改进；仿真；预测AbstractSince twenty-first Century, the social life level of people is increasing gradually, along with the vehicle in most city residents in the popular, city traffic congestion problem deteriorated rapidly. In order to solve this problem, arrange the transportation operation rationally, improve the efficiency of the use of transportation facilities, optimize the allocation of a reasonable allocation to traffic flow and signal lamp, accurate traffic flow forecasting is an important basis for solving these problems.In recent years, the BP neural network has been widely used in the prediction of. This paper designs the short-term traffic flow forecasting for the three layer BP neural network. The prediction of traffic volume in three days time recently as the input signal, the traffic volume forecast at a time. With Jiangmen urban traffic volume data based on the simulation results, the results show that the method for short-term traffic flow forecasting is feasible. Finally, the accuracy of the prediction results are analyzed and discussed.Key words Short-term traffic flow BP neural network Improve Simulation Forecast目 录摘 要.......................................................................
