- In rural areas, shared mobility-on-demand services can improve the sustainability of public transport. However, bundling customer rides is challenging due to an unfavorable spatial and temporal demand distribution. As one potential solution, service providers could apply demand management. By controlling the availability of offered rides on an operational level, they could try to influence the resulting orders to allow more bundling. In practice, however, the introduction of demand management, which is a strategic decision, is often impeded by the inability of stakeholders to assess the exact impact on system performance in advance. In this paper, we tackle this issue by developing a methodology that serves as a basis for the strategic decision on how to implement operational demand management by realizing different types of demand control policies. More precisely, we propose a methodology that evaluates different policies by applying them to a model of the operational planningIn rural areas, shared mobility-on-demand services can improve the sustainability of public transport. However, bundling customer rides is challenging due to an unfavorable spatial and temporal demand distribution. As one potential solution, service providers could apply demand management. By controlling the availability of offered rides on an operational level, they could try to influence the resulting orders to allow more bundling. In practice, however, the introduction of demand management, which is a strategic decision, is often impeded by the inability of stakeholders to assess the exact impact on system performance in advance. In this paper, we tackle this issue by developing a methodology that serves as a basis for the strategic decision on how to implement operational demand management by realizing different types of demand control policies. More precisely, we propose a methodology that evaluates different policies by applying them to a model of the operational planning problem, which itself has not been considered in the existing literature. For this purpose, we first formulate the operational planning problem as a Markov decision process. Second, we apply practical solution algorithms representing different control policies on a model variant supporting the strategic decision. Finally, drawing on real-world data from FLEXIBUS, a rural provider in Germany, we conduct a computational study and present managerial insights into the impact of different control policies on the system performance in terms of profit, which the provider aims at maximizing, and other sustainability-oriented objectives of municipal contracting authorities.…
