考虑时间阈值的铁路运输链分担率模型

以义乌-宁波国际集装箱内陆段托运人为研究对象，选取运输链属性、托运人属性和货物属性构建运输链效用函数，通过意愿调查获取托运人货运选择数据以分析其选择机理.比较了MNL 模型（Multinomial Logit）与MNLS 模型（Swait 提出的MNL）、 CMNL模型（Constrained MNL）等阈值模型在参数估计结果和选择行为解释能力上的差异.结果表明：运输链时间、费用、时间阈值和企业规模等因素对货运服务选择至关重要； 引入时间阈值可有效提高模型精度，忽视时间阈值将导致时间价值有偏；当铁路运输链 时间与时间阈值要求差距甚大或较小时，忽视时间阈值易低估或高估铁路运输链服务水 平改善所提升的分担率.

Railway Transportation Chain Share Ratio Model Incorporating Time Threshold

Shippers of international container inland segment from Yiwu to Ningbo are taken as research object, attributes of transportation chains, shippers and cargoes are then selected to construct transportation chain utility function, after that, shippers’freight choice data which are collected by stated preference investigation are used to analyze choice mechanism. MNL model (Multinomial Logit) and threshold models including MNLS model (MNL put forwarded by Swait) and CMNL model (Constrained MNL) are compared on parameter estimation results and interpretation capability of choice behavior. Results show that transportation chain time, cost, time threshold and enterprise scale of shipper play an critical role in choosing freight service, meanwhile introducing time threshold can improve the accuracy of model effectively, while ignoring the time threshold will lead to biased value of time; when railway transportation chain time is far from or near to time threshold requirement, ignoring time threshold will underestimate or overestimate the share ratio increment of railway transportation chain by improving its service.