基于ALNS-TS算法的半灵活型需求响应公交调度问题  

Semi-flexible demand responsive transit scheduling based on ALNS-TS algorithm

作　　者：靳文舟[1] 杜昊 巫威眺[1] JIN Wenzhou;DU Hao;WU Weitiao(School of Civil Engineering and Transportation,South China University of Technology,Guangzhou 510641,Guangdong Province,P.R.China)

机构地区：[1]华南理工大学土木与交通学院,广东广州510641

出　　处：《深圳大学学报（理工版）》2023年第4期425-434,共10页Journal of Shenzhen University(Science and Engineering)

基　　金：国家自然科学基金资助项目(52072128,72071079,52272310);广东省基础与应用基础研究基金资助项目(2023A1515011696)。

摘　　要：半灵活型需求响应公交模式往往拥有1个基准线路方向与1批确定访问的固定站点,服务车辆可以根据预约请求灵活响应候选的可变站点为其提供服务,当可变站点无请求时则不服务.基于低客流区域内需求响应公交服务所面临载客率低、运营成本高等问题,完善该模式的发车条件,设计差异化票价策略和未预约乘客服务判断的半灵活型需求响应公交服务系统,并建立车辆路径优化模型.为提高算法的搜索空间范围和局部搜索强度,设计了混合自适应大邻域禁忌搜索算法进行模型求解.以中国广州市黄埔区公交站点网络作为案例分析,结果表明,在新策略模式下,去除和未去除未预约乘客的半灵活型需求响应公交的票价收入较传统模式分别提高3.0%和15.5%,响应未预约乘客服务模式的服务人数可以提高7.5%;混合自适应大邻域搜索算法在求解速度和求解质量上均有较好表现,与混合模拟退火算法和遗传算法相比,计算时间分别节约76.9%和97.0%.The semi-flexible demand responsive transit is an innovative combination of traditional fixed-route transit and demand responsive transit.It is considered as semi-flexible because it has a predetermined base route and a batch of fixed stations that are determined to be visited,and has some candidate variable stations that the service vehicles can flexibly respond to provide services based on reservation requests.Responding to the problems of low passenger capacity and high operating costs faced by demand responsive transit services in low-demand areas,the departure conditions of this mode are established.A semi-flexible demand responsive transit system based on differentiated fare strategy is designed and the transit scheduling model is established.To improve the search space range and local search intensity of the algorithm,a hybrid adaptive larger neighborhood search-tabu search(ALNSTS)algorithm is designed to solve the model.The conclusion shows that under the new strategy,the semi-flexible demand responsive transits with and without reservation passengers have increased ticket fare revenue by 3.0%and 15.5%,respectively,compared with the traditional mode,and the mode with responding to unreserved passengers can increase the number of passengers served by 7.5%.ALNS-TS algorithm has better performance in solving speed and quality.Compared with simulated annealing-tabu search(SA-TS)algorithm and genetic algorithm,the calculation time of ALNS-TS is saved by 76.9%and 97.0%of the computing time,respectively.

关 键 词：交通运输工程 公共交通 需求响应公交 票价制定 路径规划 混合大邻域搜索算法 

分 类 号：U491[交通运输工程—交通运输规划与管理] C935[交通运输工程—道路与铁道工程]