The recent technological innovations in various ICT platforms have given rise to innovative mobility solutions. Such systems could potentially address some of the inherent shortcomings of a line/schedule based public transport system. Previous studies either assumed that flexible on-demand services are used as an exclusive door-to-door service or offered as a feeder connection to high-capacity public transport services. However, users may combine line/schedule based public transport systems (Fixed PT) and on-demand services (Flexible PT) so that their travel impedance is minimized. To this end, we propose a multimodal route choice and assignment model that allows users combining Fixed and Flexible PT or use them as individual modes while demand for these services is endogenously determined. The model takes into account the dynamic demand–supply interaction using an iterative learning framework. Flexible public transport can be used to perform any part of the trip, ranging from a first/last mile service to an exclusive direct door-to-door connection. The developed model is implemented in an agent based simulation framework. The model is applied to a network centered around the city of Amsterdam, The Netherlands. Scenarios where Fixed PT and Flexible PT are offered as mutually exclusive modes or can be combined into a single journey, are analysed. Results indicate that Flexible PT is predominantly used for covering $<$30% of the trip length, indicating that it is mainly used as an access or egress mode to Fixed PT. This results with an overall increase in the share of public transport trips. Also, the average waiting time of Flexible PT users when used in combination with Fixed PT are lower than the scenario where each of them is used as an exclusive mode.

各种ICT平台中的最新技术创新催生了创新的移动解决方案。这样的系统可以潜在地解决基于线路/时间表的公共交通系统的一些固有缺陷。先前的研究要么假设将灵活的按需服务用作专有的门到门服务，要么作为与大容量公共交通服务的馈线连接。但是，用户可以将基于线路/时间表的公共交通系统（固定PT）和按需服务（灵活PT）结合使用，以使旅行阻抗最小化。为此，我们提出了一种多式联运的路线选择和分配模型，该模型允许用户结合固定和灵活PT或将它们用作单独的模式，同时内生确定对这些服务的需求。该模型考虑了使用迭代学习框架的动态需求-供应互动。灵活的公共交通可用于执行旅程的任何部分，从第一英里/最后一英里服务到专有的直接门对门连接。开发的模型在基于代理的仿真框架中实现。该模型被应用于以荷兰阿姆斯特丹市为中心的网络。分析了固定PT和灵活PT作为互斥模式提供或可以组合为单个旅程的方案。结果表明Flexible PT主要用于覆盖 开发的模型在基于代理的仿真框架中实现。该模型被应用于以荷兰阿姆斯特丹市为中心的网络。分析了固定PT和灵活PT作为互斥模式提供或可以组合为单个旅程的方案。结果表明Flexible PT主要用于覆盖 开发的模型在基于代理的仿真框架中实现。该模型被应用于以荷兰阿姆斯特丹市为中心的网络。分析了固定PT和灵活PT作为互斥模式提供或可以组合为单个旅程的方案。结果表明Flexible PT主要用于覆盖$<$行程长度的30％，表明它主要用作固定PT的访问或出口模式。这导致公共交通出行比例整体增加。此外，与固定PT结合使用时，Flexible PT用户的平均等待时间比将每个人都用作独占模式的情况要短。