The Agent-based Model's Role in Assessing Paratransit Performance in a Data-Scarcity Context.

Abstract The study of informal mobility, particularly Paratransit, faces significant challenges in academic research and urban planning practices. This difficulty stems from a lack of information and resources coupled with an organized, complex nature of the Paratransit in developing countries. Making a comprehensive study on Paratransit is a resource-consuming process. This paper addresses these challenges by complementing the recent studies on paratransit modeling by proposing a pipeline to model the Paratransit corridors with multiple stops in Alexandria, Egypt. The proposed methodology aims to integrate the Paratransit into a multi-agent-based Simulation model. The process includes collecting GPS data through minibus users, constructing an Origin-Destination Matrix, and constructing, mapping, and ground-truth the Paratransit routes for an agent-based business-as-usual model. The findings highlight the variable demand for paratransit services under different mobility conditions. Ultimately, the Model was tested to make the Paratransit more accessible. The result shows a gradual change in the boarding behavior at the start and end line, while the demand fluctuated in the middle of the route. These results can support an evidence-based urban mobility policy that places a Paratransit line along the Minibus corridor. The process is based on open access data with a small sample of crowdsourcing data, enabling replicating the methods in similar data-scarcity contexts.