摘要
为了有效地定制公交线路方案以提高运行效率,针对目前定制公交多停车场多车线路优化大多采用先聚类后求解的问题,以及在进行定制公交线路优化建模时忽略上车区域到下车区域距离,或者将其设定为定值的问题,提出一种基于遗传算法的采用三段式混合编码方式的优化求解方法。根据实际过程中定制公交线路优化问题的描述,以路网中所有定制公交车辆总运营里程最小为优化目标,构建满足多个停车场、多个上下车站点、多辆定制公交车的线路优化模型。通过对模型的结构进行分析,采用包括停车场段、上车站点段、下车站点段的三段式混合编码、分段交叉以及翻转变异等遗传操作方法求解。以兰州市城关区部分交通网络为例,求解包含2个定制公交停车场、12个上下车站点的实际算例,以验证模型及算法的合理性。结果表明,采用基于遗传算法的三段式混合编码方式的算法能快速完整地求解出定制公交线路优化方案。该算法与K-means和遗传算法的混合算法相比,总运营里程减少2km,上座率提升18.375%,定制公交车辆数减少1辆,运算时间能节省38.24%。
To effectively establish route schemes of customized bus can improve operational efficiency.A threestep hybrid coding method which based on genetic algorithm is proposed to solve problems in optimization of customized bus,which are clustering at first and then solving for multi-parking areas and multi-car route and ignoring distance between pick-up areas and drop-off areas,or setting the distance as a fixed value in a process of optimization.According to the problems in actual processes,the objective is minimizing total operating mileage of all customized buses in a road network.A model for route optimization is developed to meet the needs of multiple parking lots,multiple pick-up and dropoff stations,and multiple customized buses.Genetic operation methods are used to solve the problems,including threestep hybrid coding for parking lot,pick-up station,and drop-off station,piecewise crossover,and flip mutation.Taking apart of the traffic network in Chengguan District of Lanzhou City as a case study,apractical example including 2 customized parking lots and 12 pick-up/drop-off stations are calculated to verify rationality of the model and the algorithm.The results show that the three-step hybrid coding algorithm can quickly and completely solve the optimization routes schemes of customized buses.Compared with a hybrid algorithm of K-means and a genetic algorithm,total operating mileage is decreased by 2 km;attendance rate is increased by 18.375%;number of customized buses is reduced by 1;operation time is decreased by 38.24%.
作者
王超
马昌喜
WANG Chao;MA Changxi(School of Traffic and Transportation,Lanzhou Jiaotong University,Lanzhou 730070,China)
出处
《交通信息与安全》
CSCD
北大核心
2019年第3期109-117,127,共10页
Journal of Transport Information and Safety
基金
国家自然科学基金项目(71861023)资助
关键词
交通规划
定制公交
线路优化
三段式混合编码
遗传算法
transportation planning
customized bus
route optimization
three-step hybrid coding
genetic algorithm
