定制公交线路优化综述

为全面回顾定制公交线路优化问题的研究进展，从优化目标、问题场景和求解算法3个方面对相关文献进行了归类分析。研究结果表明:定制公交线路的单目标优化研究主要集中在行驶时间、运营里程、运营成本、运营收益以及多种成本线性加权形成的系统总成本等方面，而多目标优化研究主要通过同时考虑运营成本、出行成本和服务质量中的2种或3种来实现；根据出发和到达站点的数量，定制公交线路优化的问题场景可分为“一对一”、“多对一”和“多对多”3种，针对停靠站点之间时间阻抗场景的研究主要集中在“静态时间阻抗”，对“动态时间阻抗”的研究较少；出行需求场景的研究也主要集中在“静态出行需求”，对于“动态出行需求场景”，一般通过两阶段优化策略进行求解；由于定制公交的线路优化问题属于一种特殊的车辆路径优化问题，精确求解算法适用于少量出行需求的分析案例，针对大规模出行需求的实际问题，一般采用启发式智能算法进行求解。未来的研究中，定制公交的线路优化需要考虑停车场设置和停靠点选择的影响，针对不同类型出行者设置特定的时间窗属性；此外，大数据背景下如何兼顾实时出行需求和运营成本约束，提供差异化的定制公交线路也将是具有挑战的研究方向。 

Review on customized bus route optimization

To comprehensively review the research progress in customized bus route optimization, the relevant literatures were classified and analyzed from three aspects including optimization objective, issue scenario, and solution algorithm. Analysis results show that researches on the single-objective optimization of customized bus routes have mainly focused on the travel time, operating mileage, operating cost, operating revenue, and total system cost formed by the linear weighting of multiple costs. However, research on the multi-objective optimization was mainly achieved by simultaneously considering two or three objectives, including the operating cost, travel cost, and service quality. According to the number of departure and arrival stations, the issue scenarios of customized bus route optimization problems can be divided into three types including one-to-one, many-to-one, and many-to-many. Research on the time impedance scenarios between different stops mainly focuses on the static time impedance, and less on the dynamic time impedance. Research on the scenario of travel demand mainly focuses on the static travel demand, and two-stage optimization strategies are generally used to solve dynamic travel demand scenarios. Since the route optimization problem of customized public transportation is a special vehicle route optimization problem, the precise solution algorithm is suitable for the analysis of small travel demand. For the practical problem of large-scale travel demand, the heuristic intelligent algorithm is generally used. In future studies, the optimization of customized bus routes needs to consider the influence of the parking yard settings, stop selection, and formulate particular time window attributes for different types of travelers. Besides, in the context of a big data environment, how to take into account real-time travel demand and operating cost constraints and provide differentiated customized bus routes will also be a challenging research direction. 5 tabs, 6 figs, 61 refs.