摘要
针对目前交通量预测过程中忽视交通变化固有规律的问题,从不同角度对交通量变化趋势进行了分析,提出符合一种纵横相似结合分布规律的组合交通量预测模型,利用交通量变化过程中相邻年间相同时期的纵向相似性和同年间相邻日期间的横向相似性求解出被预测天的横纵相似分时交通量数据,并将其与通过单天交通量分布规律求解出的分时分布交通量数据相叠加,最终实现对交通量的预测。结果表明:各阶段预测结果的平均绝对误差均在6%以内,且能够对工作日、周末以及节假日不同类型的日期进行交通量预测。
In order to solve the problem of ignoring the inherent law of traffic change in the current traffic forecasting process, the trend of traffic volume change was analyzed from different angles, and a combined traffic volume forecasting model with vertical and horizontal similarity distribution was proposed. By using the change of traffic volume the vertical similarity of the same period in the neighboring years and the horizontal similarity between the adjacent days of the same year, the vertically and horizontally similar time-sharing traffic data of the predicted days were calculated and compared with the time-sharing distribution data obtained by the single-day traffic distribution rule, and finally the forecast of traffic volume was achieved. The experimental results show that the average absolute error of the predicted results is less than 6%, and it is possible to forecast the traffic volume for different typesof working days, weekends and holidays.
出处
《计算机应用》
CSCD
北大核心
2017年第A02期68-73,共6页
journal of Computer Applications
基金
吉林省产业技术研究与开发专项(2016C090)
吉林省交通运输厅应用系统开发类项目(2015-1-22)
关键词
交通工程
纵横相似
分布规律
高速公路
交通量预测
traffic engineering
vertical and horizontal similarity
distribution
expressway
traffic flow prediction
