Visiting Fellowship for Juan de Dios Ortuzar

旅行需求模型用于估计人口旅行需求和旅行网络提供的变化对旅行的数量和性质的可能影响，特别是对每种可用方式进行的旅行的影响。大多数旅行需求模型是基于横断面数据，即在同一时间观察到的数据。然而，这种类型的数据通常不能提供有关选择的令人满意的信息(即使在特定时间)，因为不可能知道它们将如何随系统变化而变化。此外，很难考虑之前选择的影响。如果我们对估计未来的交通需求感兴趣，这一限制尤其重要，这是城市地区长期规划中的一个关键方面。面板数据是一个很好的替代方案，分析师在不同的时间反复采访一组人，因为面板是专门为纳入时间影响而设计的。然而，尽管自Heckman的开创性工作以来，面板数据在许多领域的使用有所增加，但在交通规划方面的应用很少，这些应用可分为两类：-长期调查小组，即相同的调查(即相同的设计和方法)每年重复一次或两次，持续一定年限，或在重要事件之前和之后重复。这里的主要问题是连续小组浪潮之间的损耗(即失去受访者)。-短期调查小组：由多天数据组成，其中在连续两天或更多天收集同一样本的重复测量数据，但在随后的几年中不重复进行调查。最近的例子是美国收入动态国家小组研究的两天时间使用日记，以及德国和瑞士为期六周的旅行和活动日记小组。虽然自然损耗在这里不是问题，但模式选择的不频繁变化和低数据变异性(模式属性和个体的社会经济特征实际上是固定的)可能会导致估计模型的困难。当在系统中引入实质性干预时，数据面板在评估变化方面具有更显著的优势。事实上，政策的影响可能会改变趋势，使用对相同个人的观察可以更容易地捕捉到这些变化(因为当前的部分行为可以用以前的经验来解释)。尽管小组的优势看起来很明显，但我们并不知道围绕系统的实质性变化而计划的经验，以便估计模式选择过程的改进模型。两个值得注意的例外是阿姆斯特丹围绕其城市高速公路系统的延伸而开发的工作和最近在智利完成的圣地亚哥小组，该小组是围绕其全新的公共交通系统--仍然臭名昭著的Transantiago--的实施而设计的，目的是制定和估计包含惯性和政策冲击效应的模型。由于小组调查等新技术的迅速发展，以解决交通规划中的重要问题，继续需要最新的文本，就其使用提供明确和合理的指导。此外，在重要的专业会议上发表最先进的论文将加速将得到适当支持的小组调查转化为实践。此次访问的目的是为制定和评估改进的模式选择模型作出重大贡献，以便能够考虑惯性和政策冲击的影响，并将这些技术纳入广泛可用的参考和教材中。

Travel demand models are used to estimate the likely effect of changes in the need in the population for travel and in the provision in networks for travel on the amount and nature of travel that is undertaken, and in particular on the travel that is undertaken by each of the available modes.The majority of travel demand models are based on cross-sectional data, that is, data observed at a single time. However, this type of data does not generally yield satisfactory information about choices (even at a particular time), because it is not possible to know how they would vary in response to system changes. Also, it is difficult to consider the influence of previous choices. This limitation is particularly significant if we are interested in estimating future transport demand, a key aspect in for the long-term planning of urban areas.Panel data, where the analyst interviews repeatedly a group of individuals at separate times, are a good alternative as panels are designed specifically to incorporate temporal effects. However, and although the use of panel data has increased in many areas since the pioneering work of Heckman, in transport planning there are few applications, and these can be classified in two groups:- Long survey panels, where the same survey (i.e. same design and methodology) is repeated once or twice a year for a certain number of years, or before-and-after an important event. The main problem here is attrition (i.e. loosing respondents) between successive panel waves.- Short survey panels: which consist of multi-day data where repeated measurements on the same sample are gathered over two or more successive days, but the survey is not repeated in subsequent years. Recent examples are two-day time-use diary for the US National Panel Study of Income Dynamics and six-week travel and activity diary panels in Germany and Switzerland. Although attrition is not a problem here, the infrequent changes in mode choice and low data variability (the modal attributes and the individuals socioeconomic characteristics are practically fixed) may cause difficulties in estimating models.When a substantive intervention is introduced in a system, data panels have even more significant advantages for evaluating changes. Indeed, the effects of policies could change trends, and it can be easier to capture these changes using observations for the same individuals (because part of the current behaviour could be explained by previous experience).Although the advantages of panels seem clear, we are not aware of experiences being planned around a substantial change in the system in order to estimate improved models of the mode choice process. Two notable exceptions are the work developed in Amsterdam around an extension of its urban motorway system and the recently completed Santiago Panel in Chile, which was designed around the implementation of its radically new public transport system, the still infamous Transantiago, in order to formulate and estimate models incorporating both inertia and policy shock effects.Because of the rapid development of novel techniques such as panel-based surveys to address important issues in transport planning, there is a continuing need for up to date texts providing clear and sound guidance on their use. Furthermore, the presentation of state-of-the art papers at important professional conferences will expedite the transfer of properly supported panel surveys into practice.The aims of this visit are to make a significant contribution to the establishment of good practice for the formulation and estimation of improved mode choice models, allowing to consider the effects of inertia and policy shocks, and to incorporate these techniques into widely available reference and teaching material.