校车周末高峰时段交通量预测与特殊日期调整

School Bus Peak Period Traffic Volume Forecast-ing in Weekend and Special Day Adjustment

准确的校区间高峰时段交通量预测,不仅可以帮助校车运营部门在高峰时段确定合理的发车间隔提高运营效率节约运营成本;还可以减少师生的候车时间,改善校车服务水平减少拥挤保证运行安全都有重要意义。采用BP神经网络模型对校区间周末高峰时段交通量进行了预测,选择了温度和下雨量2个因素作为输入变量,在训练数据的选取上采用了移动窗口的方法。另外模型提供了对特殊日期运量预测的调整。通过对实测数据进行验证,结果表明该模型简单、快捷和实用。

Accurate forecasting of school bus peak period traffic volume in weekend,not only can help the school bus director to arrange the interval of buses to improve the efficiency and save the cost,but also can reduce the teachers and students long waiting time and improve school bus＇s service level.There are also important meanings in guarantee the school bus＇s safety without crowding.This paper presents a BP neural network model for peak period traffic volume forecasting of weekend between different school regions.Temperature and rain volume are selected as two input variables.The moving window data selection strategy was adapted in training data selection. Moreover,the model presents a unique adjustment method to adapt the special days.The result indicates that the model is simple,fast,and applicable.