ABSTRACT

This paper proposes a bi-objective multi-period bus headway setting problem while considering the influence of electronic bus station boards on passengers’ choice behaviour. A non-dominated sorting genetic algorithm II (NSGAII) based local search algorithm is developed to find the approximate Pareto front in terms of two objectives (i.e., expected average travel time of passengers and fleet size). A bus network simulation model is built to capture the complex interactions among bus stops, vehicles, and passengers. Experiment results conducted on Mandl’s network show that the proposed solution method is more capable of adapting the random passenger arrivals. The comparison of solutions with and without electronic bus station boards suggests that (i) the electronic bus station boards do not always exert a positive effect on passenger travel time; (ii) and the bus arrival information can positively affect the route choice behaviour of passengers only when the fleet size is relatively large.

摘要

本文在考虑电子公交站牌对乘客选择行为的影响的情况下，提出了一个双目标多周期公交车头时距设置问题。开发了一种基于非支配排序遗传算法II（NSGAII）的局部搜索算法，以根据两个目标（即乘客的预期平均旅行时间和车队规模）找到近似的帕累托前沿。建立了一个公交网络仿真模型来捕捉公交车站、车辆和乘客之间的复杂交互。在 Mandl 网络上进行的实验结果表明，所提出的求解方法更能适应随机到达的乘客。有和没有电子车站牌的解决方案的比较表明：(i) 电子车站牌并不总是对乘客旅行时间产生积极影响；