Abstract: Urban public transit network consists of subway lines, light-rail lines, special bus lanes, and general bus lines, and different travel speed and transfer time can be obtained by these lines. Because of the obvious randomness characters of the travel time of public transit, the optimum route search based on static traffic data is far deviated from the real one. Taking the trip sectors such as waiting, transfer, on board as links, the multi-level public transit network containing subway lines, light-rail lines, special bus lanes, and general bus lines is established in this paper. Considering the random fluctuation of travel time, it takes the average travel time and travel time reliability as link weights. By heuristic increasing link travel time weight during the path search procedure, high delay risk links are effectively avoided and the reliable path under the constraint of circuitous could be obtained. The proposed algorithm is proved to be reasonable and effective at the final part of the paper.

Key words: reliability, multi-level, public transit, path search, heuristic algorithm

摘要： 城市公交系统包含多种公交模式如地铁、轻轨、公交专用道、地面公交等，不同模式的车辆行驶速度及换乘时间差异很大，同时由于城市公交运行时间的具有较强的随机性，基于静态分析的路径选择与实际的动态最佳路径常有较大偏离。本文以公交出行链各环节为边，通过构建包含不同公交模式的多层次网络邻接关系，并对各边进行平均出行时间及时间可靠度的双重赋权，建立考虑出行各环节时间随机波动的多模式公交线网；并以通行时间较短、延误风险较小为路径寻优的双重目标，通过在路径寻优过程中对延误高发线路及换乘点进行启发式规避，实现在出行时间许可范围内有效减少延误风险的可靠路径的快速搜索。文中最后通过实例表明该算法的合理性及有效性。

关键词: 可靠性, 多层次, 公交, 路径搜索, 启发式算法