摘要
滞站调度策略是公交日常运营中最常用的一种控制策略.针对传统滞站策略存在较高误控率的问题,提出一种新型的协控准点滞站调度策略,该策略依据车辆在当前站点和下一站点的准点信息,来综合判断是否需要对该车辆进行滞站控制.为了获得车辆在下一站点的发车时间信息,设计了基于支持向量机的公交旅行时问预测模型.最后,利用基于Paramics的仿真实例对本文提出的预测模型和调度策略进行了验证,结果表明,基于支持向量机的预测模型具有较高的预测精度,可以为协控滞站调度策略提供较可靠的依据;协控准点滞站调度策略比传统的滞站调度策略具有更低的误控率和较少的乘客等待费用.
Holding strategy is one of the most commonly used operation control strategies. This paper presents an improved schedule-based holding strategy, in which a support vector machine （SVM）-based model for fore- casting the early bus departure time from the following stop, is developed. This strategy judges whether an early bus should be held according to the on-time information of the bus at the current stop and the following stop. Finally, a simulation-based case is used to illustrate the performance of SVM-based model and the im- proved holding strategy. Results show that the SVM-based model can achieve accurate prediction and provide the improved holding strategy with reliable information. It is shown that the improved holding strategies show better performance than the traditional schedule-based holding strategies.
出处
《系统工程学报》
CSCD
北大核心
2012年第2期248-255,共8页
Journal of Systems Engineering
基金
辽宁省百千万人才工程基金资助项目(2008921081)
国家自然科学基金资助项目(51108053)
关键词
公交调度
滞站调度策略
支持向量机
bus operation
coordinated schedule-based holding strategy
support vector machine（SVM）
