A scaled and sustainable demand responsive transport service

由于车辆的制造、使用、储存和处置，私人机动性的碳足迹很高。私家车有96%的时间闲置，占道路运输二氧化碳排放总量的60.7%。为了在减少二氧化碳排放的同时减少社会损失，将服务水平较低的地区联系起来，并减轻老年人和残疾人在支付出行费用方面的挑战，本研究建议开发提供可持续、共享出行所需的数学工具，特别是需求响应式交通服务(DRTS)。我们将设计新的算法来优化与动态定价相结合的DRTS的路由和调度。实时解决这些大规模的硬组合优化问题，将使DRTS实现转型，DRTS是规模化共享交通解决方案新兴领域的一部分，鼓励更多人接受共享交通。DRTS允许乘客预订门到门服务，要求上门或下车时间，这很像出租车，但与沿途可能上下车的其他乘客共用一辆车。类似的服务，如Dial-a-Ride，是为了满足特定需求而存在的，但它们的范围有所缩小，并得到地方议会和交通部的大量补贴。它们缺乏路线规划的灵活性，无法应对高需求。在规模上，通过优化动态定价和路线、现实的需求预测、知情的准确行为模型，以及提高其接受度和鼓励自愿行为改变的政策的激励，DRTS在财务上将是可行的，比私家车使用更可持续。该项目以高效、复杂的优化算法为形式的原创变革性科学，以及对共享出行偏好和态度的丰富理解，将有助于使DRTS既高效又经济；从而促进共享出行并显著减少当地出行的二氧化碳排放。该项目将整合三个重要的科学组成部分，以提供有吸引力的、灵活的、低碳的DRTS。1)针对不同类型车辆的车队的有效、高效的调度和路线优化算法，该算法可以针对行程请求提供即时的接受/拒绝决策。为了有效地做到这一点，该算法需要预测潜在的未来需求，并在后台持续地对整个机队的调度进行全局优化。2)新的收入管理公式，允许行程价格动态变化，价格取决于行程长度和服务质量；因此，支持服务的财务可持续性。3)对客户行为和偏好有丰富的了解，这将通过开展调查和焦点小组，并使用收集的数据来建立选择模型，描述潜在乘客如何做出决定来获得。这些模型将支持服务设计和激励行为改变。将这三个工作组成部分结合起来，全面解决实际挑战，并推进共享绿色交通的一个令人兴奋的跨学科研究领域。这种算法方法也有可能在未来适用于电动和自动驾驶汽车。

Private mobility has a high carbon footprint due to the manufacturing, use, storage and disposal of vehicles. Private cars spend 96% of their time idle and were responsible for 60.7% of total CO2 emissions from road transport. To reduce CO2 emissions while mitigating societal loss, linking poorly served geographies and alleviating the challenges of elderly and disabled to afford mobility, this research proposes the development of the mathematical tools needed to deliver sustainable, shared mobility, specifically a Demand Responsive Transport Service (DRTS). We will design novel algorithms that optimise the routing and scheduling integrated with dynamic pricing of DRTS. Solving these large-scale hard combinatorial optimisation problems, in real time, will enable a transformation of DRTS, part of the emerging sector of scaled shared transport solutions, encouraging increased take up of shared mobility. DRTS allows passengers to book a door-to-door service requesting pick up or drop off times, much like a taxi, but sharing a vehicle with other passengers that may be collected or dropped off along the route. Similar services, such as Dial-a-Ride, exist to meet specific needs but they are reduced in scope and heavily subsidized by local councils and the Department for Transport. They lack route planning flexibility and cannot manage high demand. At scale, with optimized dynamic pricing and routing, realistic demand forecasts, informed accurate behavioural models, and incentivised by policies that enhance their acceptance and induce voluntary behaviour changes, DRTS would be financially viable and more sustainable than private car use. The original transformative science in the form of efficient, complex optimization algorithms, and the rich understanding of preferences and attitudes towards shared mobility developed in this project will help enable DRTS to be both efficient and cost-effective; thus, promoting shared mobility and significantly reducing CO2 emission of local travel.This project will integrate three important scientific components to deliver an attractive, flexible, low-carbon DRTS. 1) An effective efficient scheduling and routing optimisation algorithm for a fleet of vehicles of different types that can provide instant accept/reject decisions on journey requests. In order to do this effectively, the algorithm needs to anticipate potential future demand and be continuously globally optimising schedules across the fleet in the background. 2) New revenue management formulations that allow the prices of journeys to be changed dynamically, with prices dependent on journey length and service quality; thus, supporting the financial sustainability of the service. 3) A rich understanding of customer behaviour and preferences, which will be obtained by running surveys and focus groups and using the data collected to build choice models, describing how potential passengers make decisions. These models will support service design and motivate behaviour changes.Combining these three components of work comprehensively addresses the practical challenge and advances an exciting new interdisciplinary research area for shared green transportation. The algorithmic approach also has the potential to be adapted to electric and autonomous vehicles in the future.