面向运营管理的多粒度省际客运需求预测方法

客运量是表征省际客运需求,开展行业运营管理的基础指标.为增强省际客运行业运营管理水平,提高旅客出行效率和应急保障能力,建立面向省际客运运营管理的年客运量和节假日客运量的多粒度预测模型.在影响因素与省际年客运量关联度分析的基础上,构建基于BP神经网络的年客运量预测模型.考虑特殊节假日的影响特征,提出了指数平滑与季节模型相结合的节假日客运量组合预测模型,实现节假日总客运量、日客运量的预测.以北京的实际数据为例,对预测模型进行精度验证.结果表明,年客运量预测模型的平均相对误差为0.15%,春运期间每日客运量预测模型的平均相对误差为6.7%,能较好地体现客运量在不同阶段的变化趋势,具有良好的稳定性. 

A Method of Multi Granularity Intercity Passenger Demand Forecasting for Operation Management

Passenger volume is a basic indicator of the demand for inter-provincial passenger transport,which reflects levels of operation and management in this industry.A forecasting model with multi-granularity for passengers' volume in one year or on holidays is developed to promote level of management,travel efficiency of passengers,and capacity of emergency response.A prediction model of passengers' volume using BP neural network based on correlation analysis of influencing factors and annual passengers' volume is developed.Considering special influencing characteristics of passengers' volume on holidays,a forecasting model combines an exponential smoothing model and a seasonal model is proposed.The total and daily volume of inter-provincial passenger transport during holidays is predicted.Taking actual transport data in Beijing as a case study,the accuracy of the prediction model is verified.The results show that the average relative error of the prediction model of annual passenger volume is 0.15%,and the average relative error of the prediction model of daily passenger volume during the Spring Festival is 6.70%.These indicating that the prediction models can reflect the variation trend of passenger volume in different periods,and has good stability.