多换乘点响应型接驳公交运行线路的协调优化

Coordinated Optimization of Operation Routes for Responsive Feeder Transit Systems with Multiple Transfer Points

为尽量降低响应型接驳公交系统的运行费用,提出多换乘点间运行线路协调设计的构想.针对同时包含预约需求和实时需求的混合需求,构建多换乘点响应型接驳公交系统运行线路的2阶段协调优化方法,并设计优化流程.第1阶段仅考虑预约需求,首先将预约乘客按有/无特定换乘点要求进行分类,在此基础上构建预约需求下多换乘点多车辆运行线路的协调优化模型.在协调优化模型中,优化目标是由乘客时间费用、车辆运行费用、以及惩罚费用所构成的系统总成本最小;乘客时间费用包括乘客候车时间的惩罚费用、车内乘客在需求点的等待时间费用以及乘客车上时间的惩罚费用3个部分;车辆运行费用包括车辆启动费用、路段行驶费用、需求点的停靠费用、车辆早到引起的等待费用4个部分;考虑的约束条件包括乘客候车和车上的软时间窗、乘客换乘点要求、车辆容量、车辆出行时长等.第2阶段根据规则判断是否响应实时需求,并根据响应情况重新优化后续各班次的运行线路.针对第1阶段模型,基于模拟退火算法设计求解算法.研究表明:在预约需求或混合需求条件下,与各换乘点运行线路独自优化相比,协调优化方法均能显著降低运送全部响应乘客所需的平均运行距离和平均总成本;仅有预约需求时分别降低5.4%、19.8%,新增实时需求后分别减少1.4%、21.7%;与固定发车间隔相比,分时段调整发车间隔,也能有效降低运送全部响应乘客所需的平均运行距离和平均总成本,仅有预约需求时分别降低18.2%、17.2%,新增实时需求后分别减少19.97%、25.06%,说明多换乘点间车辆路径的协调运行是提升响应型接驳公交运行效率的有效途径.

To minimize the operational costs of responsive feeder transit systems, a coordinated design of operation routes between multiple transfer points was proposed in this study. A two-stage coordinated optimization method for operating routes of responsive feeder transit systems with multiple transfer points under mixed demands, including reserved and real-time demands, was constructed, and the optimization process was designed. In the first stage, only the reserved passengers were considered. First, the reserved passengers were classified based on the requirements of specific transfer points. Then, the coordinated optimization model under the reserved demands for the operation routes of multiple vehicles and multiple transfer points was constructed. In this model, the optimization goal was to minimize the total cost of the system, which comprised costs related to passenger time, vehicle operation, and penalty. The passenger time cost was divided into three parts:penalty of wait time, wait time of the passenger on-the-vehicle at the demand point, and penalty of on-the-vehicle time. Similarly, the vehicle operational cost was divided into four parts:start-up cost, road travel cost, stop cost at the demand point, and wait time cost caused by the early arrival. The soft time window of passenger waiting and on-the-vehicle, the requirements of passenger transfer points, vehicle capacity, and travel time were considered as constraint conditions. In the second stage, the decision on whether to respond to real-time demand was made based on the rules. In addition, the operation routes of subsequent shifts were re-optimized according to the response situation. For the first-stage model, the solution algorithm based on Simulated Annealing algorithm was designed. The example showed that under the conditions of reserved or mixed demand, compared with the separate optimization for the operation route of each transfer point, the average running distance and average total cost required for transporting passengers could be significantly reduced by the coordinated optimization method proposed in this study. Specifically, they are reduced by 5.4% and 19.8%, respectively, under the reserved demand. After real-time demand was added, they are reduced by 1.4% and 21.7%, respectively. Compared with the fixed departure interval, the average running distance and average total cost of transporting passengers can also be effectively reduced by adjusting the departure interval in different periods. Specifically, they are reduced by 18.2% and 17.2%, respectively, under reserved demand. After real-time demand was added, they are reduced by 19.97% and 25.06%, respectively. This shows that the coordinated operation of the vehicle route between multiple transfer points is an effective means of improving the efficiency of responsive feeder transit systems.