基于双种群遗传算法的公交线路发车间隔优化

A bus headway optimization model with dual-population genetic algorithm

提出一个公交线路发车间隔优化模型,以公交系统的社会总效益最大化为目标,兼顾乘客和运营者双方利益,设计公交网络中各条线路的发车间隔.该模型在车辆资源不变的约束下,通过线性加权法衡量乘客和运营者的利益,以达到系统最优目的.为求解该模型,开发了一个双种群的遗传算法,该算法可有效保持遗传算法进化过程中的多样性,提高优化质量.以大连市主城区公交系统的数据对该模型和算法进行检验,结果表明,若整合大连市公交车辆资源,可改善整个系统的服务水平,且降低系统总成本.

A model for bus headway optimization is presented, aiming to minimize the overall cost of passengers and the bus operator, considering the vehicle fleet constrain. The cost the parties concerned can be measured by a linear weighted technique. A dual-population genetic algorithm is proposed to solve the headway optimization model. This model can keep the diversity of this algorithm during its evolution, which will greatly improve the performance of the genetic algorithm. Finally, data collected in Dalian city, China, are used to verify the model and algorithm. Results show that reasonable resource assessment can improve the service quality and decrease the cost of the transit system.