Abstract:

The trip time (ST), as an essential parameter, greatly affects the compilation and on-time probability of a vehicle schedule in public transit services. Unfortunately, the significance of ST has not been generally recognized in China. The values of STs are usually set manually based on experiences, which are normally hard to reflect the realworld situation. To set proper trip times, a novel automatic approach based on the AVL data is proposed. The basic process is as follows. First, running time samples are abstracted from a large set of AVL data, based on which the homogeneous running time (HRT) bands are then decided. Meanwhile, the knowledge base of running time distribution is established. Then, the ST parameters corresponding to each HRT band are generated based on a waiting time model. Finally, a simulation system is developed to test the schedules compiled based on a given set of STs, which may be revised further according to the simulation results. Experiments on the bus line 4 of Haikou of China show that setting the STs by the proposed approach brings high ontime probability to vehicle schedules.

Key words: intelligent transportation, trip time, AVL data analysis, heuristics, service reliability measurement

摘要：

单程时间（ST）是编制公交运营计划的重要参数，而国内公交企业对此普遍重视不足，多采用公司统一而非针对不同线路的时段标准，由调度人员凭借经验设置参数值，致使其无法反映现实状况,运营服务准点率低下.为此，本文提出一个基于车辆定位（AVL）数据的ST设置方法.其基本思路是：首先通过分析大量AVL数据中的实际单程时间样本，划分“同运营时间（HRT）时段”，并建立“单程分布知识库”；然后采用候车时间优化模型，计算各HRT时段中的ST参数取值；最后，对ST参数进行仿真验证和调整.对海口4路现实数据的实验显示，采用该方法设置ST参数有助于编制出高执行率和准点率的行车方案.

关键词: 智能交通, 单程时间, AVL数据分析, 启发式方法, 运营服务分析