Designing large-scale bus network with seasonal variations of demand
Designing large-scale bus network with seasonal variations of demand
复制标题
DOI：
10.1016/j.trc.2014.08.017
复制
发表时间：
2014-11
影响因子：
8.3
通讯作者：
S. M. Amiripour;A. Ceder;A. S. Mohaymany
中科院分区：
文献类型：
--
作者：
S. M. Amiripour;A. Ceder;A. S. Mohaymany
建立一个方便满足乘客需求的公交网络是公交运营规划过程的重要组成部分。当然，最佳巴士网络的决定对需求的任何变化都非常敏感，因此，最好不要考虑平均或估计数字，而是要谨慎考虑需求的变化。全球许多城市都经历着季节性的需求变化，这自然会对过境服务的便利性和最佳性产生影响。也就是说，公共汽车网络应该在所有季节提供方便的服务。这一问题，在这项工作中解决，没有得到彻底处理，无论是在实践中，也没有在文献中。分析季节性的公交需求变化进一步增加了公交网络设计问题的计算复杂性，这被称为NP难问题。针对这一繁琐的问题，提出了一种利用遗传算法进行求解的方法，该方法有效地解决了这一问题。开发的方法被应用到两个基准网络和案例研究，在伊朗的马什哈德市超过320万居民和2000万游客每年。案例研究的特点是一个显着的季节性需求变化，演示了如何找到最好的单一网络的巴士路线，以适应每年的乘客需求的波动。将所提出的算法与之前开发的算法进行比较的结果表明，就目标函数值而言，新开发的算法比其他方法的性能高出1%到9%。
Creating a bus network that covers passenger demand conveniently is an important ingredient of the transit operations planning process. Certainly determination of optimal bus network is highly sensitive to any change of demand, thus it is desirable not to consider average or estimated figures, but to take into account prudently the variations of the demand. Many cities worldwide experience seasonal demand variations which naturally have impact on the convenience and optimality of the transit service. That is, the bus network should provide convenient service across all seasons. This issue, addressed in this work, has not been thoroughly dealt with neither in practice nor in the literature. Analyzing seasonal transit demand variations increases further the computational complexity of the bus-network design problem which is known as a NP-hard problem. A solution procedure using genetic algorithm efficiently, with a defined objective-function to attain the optimization, is proposed to solve this cumbersome problem. The method developed is applied to two benchmarked networks and to a case study, to the city of Mashhad in Iran with over 3.2 million residents and 20 million visitors annually. The case study, characterized by a significant seasonal demand variation, demonstrates how to find the best single network of bus routes to suit the fluctuations of the annual passenger demand. The results of comparing the proposed algorithm to previously developed algorithms show that the new development outperforms the other methods between 1% and 9% in terms of the objective function values.