Abstract: The flexibility in Bus Rapid Transit scheduling and operation is the great difference with ordinary buses. In order to improve the quality of BRT operation, the headway and scheduling combination form of Bus Rapid Transit vehicles was studied. Models based on minimizing passengers’ travel costs and transit vehicles’ travel time cost models were established. In addition, three different schedules from normal, zone and express schedules were adopted. Constraints including passenger volume, time and frequency were considered. The genetic algorithm of variable-length coding was used to solve the model and the numerical case was given. The result indicates that the optimization results can educe 69.92% cost mostly, which shows that the method is scientific and feasible.

Key words: transit operation, bus rapid transit, scheduling, genetic algorithm

摘要： 快速公交的运营调度灵活性是其与普通公交区别中很重要的一个方面。为了提高快速公交的运营调度质量，研究了优化快速公交的发车间隔和车辆形式组合，通过建立乘客出行时间成本和公交车辆出行时间成本最低的参数优化模型，考虑了全程车、区间车和大站快车三种不同车辆调度形式的组合，以及在乘客人数、时间和发车频率方面的约束条件。给出了基于可变长度编码的遗传算法的模型求解算法，并给出数值算例。结果表明：优化结果最多可以节省69.92%的成本，敏感性分析表明在出行量增加并且出行速度减小的情况下，都可以通过合理的配置车辆的形式，减少整个系统的出行成本。具有良好的科学性和可操作性。

关键词: 公交运营, 快速公交, 调度, 遗传算法