出行即服务环境下地铁多模式出行替代潜力评估OA
Evaluation of the substitution potential of metro-integrated multimodal travel in the MaaS environment
[背景]一体化的出行即服务(Mobility as a Service,MaaS)平台能够实现地铁与其他交通方式的无缝衔接,显著提升地铁多模式出行的可达性与便捷性,因此有望减少出行者对小汽车的依赖,进而缓解城市交通拥堵.[目标]以网约车出行为基准,评估MaaS环境下地铁多模式出行对其的替代潜力,并识别影响替代潜力的关键因素.[方法]首先,基于MaaS平台将地铁与常规公交、共享单车、拼车等绿色出行方式进行整合,设计十类地铁多模式出行替代方案,并根据接驳距离对替代方案进行匹配;然后,提出综合考虑经济效益、低碳效益与时间成本的替代方案综合效益计算方法,以评估地铁多模式出行的替代潜力;最后,构建CatBoost机器学习模型预测不同条件下网约车出行的可替代性,并采用累积局部效应图分析其影响因素的非线性作用.[数据]基于上海市网约车出行订单数据进行实证研究.[结果]总体约56%的网约车出行可被地铁多模式出行方案所替代,其中市中心区域的可替代比例最高,可达80%以上;影响网约车出行可替代性的关键因素包括网约车出行距离、时间延误比例、替代方案类别、地铁多模式出行距离以及起终点周边建成环境.[结论]研究证实了在MaaS环境下地铁多模式出行对于减少小汽车依赖性具有较大的潜力,同时为政府进一步优化MaaS系统提供了决策依据,有助于提高城市公共交通的分担率.
[Background]An integrated mobility-as-a-service(MaaS)platform facilitates seamless metro-multimodal connections,improving accessibility and reducing reliance on private cars,there-by alleviating urban congestion.[Objective]Using ride-hailing trips as a benchmark,this study as-sesses the substitution potential of metro-integrated multimodal travel in the MaaS environment.It identifies the key factors influencing this substitution.[Method]First,based on the MaaS platform,the metro system is integrated with green travel modes such as conventional buses,shared bicycles,and ridesplitting to design ten types of metro-integrated multimodal travel substitution schemes,which are then matched to ride-hailing trips according to connection distances.Next,a comprehen-sive benefit evaluation method is proposed that incorporates economic gains,carbon-reduction bene-fits,and time costs to assess the substitution potential of metro-integrated multimodal travel.Final-ly,a CatBoost machine learning model is developed to predict the substitutability of ride-hailing trips under different conditions,and accumulated local effect(ALE)plots are used to explore the nonlinear effects of influencing factors.[Data]Empirical analysis is conducted using ride-hailing or-der data from Shanghai.[Result]The results show that approximately 56%of ride-hailing trips can be substituted by metro-integrated multimodal travel schemes,with the highest substitutability(ex-ceeding 80%)observed in central urban areas.Key factors affecting ride-hailing substitutability in-clude ride-hailing travel distance,proportion of travel delay,substitution scheme,distance of the metro-integrated multimodal trip,and built-environment characteristics around origins and destina-tions.[Conclusion]The study confirms the substantial potential of metro-integrated multimodal travel to reduce dependence on private cars in the MaaS environment.It also provides policymakers with evidence-based insights to optimize MaaS systems and increase the share of urban public trans-portation.
李文翔;刘博;袁炫渝;申锐滔;陈培焱
上海理工大学,管理学院,上海 200093四川省交通运输发展战略和规划科学研究院,成都 610041上海理工大学,管理学院,上海 200093上海理工大学,管理学院,上海 200093浙江大学,计算机科学与技术学院,杭州 310027
交通工程
城市交通出行方式替代机器学习地铁多模式出行出行即服务
urban traffictravel mode substitutionmachine learningmetro-integrated multimodal travelmobility as a service
《交通运输工程与信息学报》 2026 (2)
53-66,14
国家自然科学基金项目(72471149)教育部人文社会科学研究项目(24YJCZH147)上海市哲学社会科学规划青年课题项目(2023ECK003)上海市教育委员会"人工智能促进科研范式改革赋能学科跃升计划"专项
评论