摘要
为了更好地理解枢纽送站坪出租车落客行为,提高落客区域通行效率,提出基于Gamma混合模型的出租车落客决策模型.应用精确的出租车轨迹数据,将出租车停车时间分解为主动停车时间、被迫停车时间和落客时间,基于被迫停车时间分析构建等待耐性混合分布模型,模型验证结果与真实数据相吻合.在此基础上,以潜在乘客耐心分布、停车位置、期望停车位和行程时间为落客决策模型的核心指标,提取相关因子为解释变量,以是否落客为被解释变量,构建二元面板Logit模型,并对模型进行检验.结果表明,乘客耐心对车辆落客起着决定性的作用,落客决策模型预测准确率超过81%,表明该模型能够较好地预测出租车落客行为,为研究缓堵策略提供基础.
A taxi drop-off decision model was proposed based on the Gamma hybrid model,in order to better understand the taxi drop-off behavior and improve the drop-off area’s traffic efficiency.The accurate taxi trajectory data was used,and the taxi stopping time was decomposed into active stopping time,forced stopping time,and dropoff time.A mixed distribution model of waiting tolerance was established based on the analysis of forced stopping time.The verification results of the model were consistent with the actual data.On this basis,the potential passenger patience distribution,parking location,expected parking space,and travel time were taken as the core indicators of the drop-off decision model.Then,relevant factors were extracted as explanatory variables,with drop-off or not as explained variables.Finally,a binary panel Logit model was constructed and tested.Results show that passenger patience plays a decisive role in vehicle drop-off.The prediction accuracy of the drop-off decision model is more than 81%,which indicates that the model can well predict taxi drop-off behavior and provides a research basis for the further research on congestion reduction strategy.
作者
杨方宜
杨荣根
李伟兵
何向东
YANG Fangyi;YANG Ronggen;LI Weibing;HE Xiangdong(College of Intelligent Science and Control Engineering,Jinling Institute of Technology,Nanjing 211169,China;School of Mechanical Engineering,Nanjing University of Science and Technology,Nanjing 210094,China;Zhejiang Hooeasy Intelligent Technology Co.Ltd,Jinhua 321042,China)
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2024年第3期589-598,共10页
Journal of Zhejiang University：Engineering Science
基金
江苏省高等学校自然科学基金资助项目(22KJD580003)
金陵科技学院博士科研启动基金资助项目(jit-b-202113)
金陵科技学院科研基金孵化资助项目(jit-fhxm-202105)。
