A Data-Efficient Approach for Evacuation Demand Generation and Dissipation Prediction in Urban Rail Transit System

• Main Authors: Xiaoqing Dai, Han Qiu, Lijun Sun
• Format: Article
• Language: English
• Published: MDPI AG 2021-08-01
• Series: Sustainability
• Subjects: urban rail transit; disruption; evacuation demand; prediction; data-efficient; simulation
• Online Access: https://www.mdpi.com/2071-1050/13/17/9692

Predicting evacuation demand, including its generation and dissipation process, for urban rail transit systems under disruptions, such as line and station closure, often requires comprehensive historical data recorded under homogeneous situations. However, data under disruptions are hard to collect due to various reasons, which makes traditional methods impractical in evacuation demand prediction. To address this problem from the modeling perspective, we develop a data-efficient approach to predict evacuation demand for urban rail transit systems under disruptions. Our model-based approach mainly uses historical data obtained from the natural state, when no shocks take place. We first formulate the mathematical representation of the evacuation demand for every type of urban rail transit station. Input variables in this step are location features related to the station under the disruption, as well as an origin–destination matrix under the natural state. Then, based on these mathematical expressions, we develop a simulation system to imitate the spatio-temporal evolution of evacuation demand within the whole network under disruptions. The transport capacity drop under disruptions is used to describe the disruption situation. Several typical scenarios from the Shanghai metro network are used as examples to implement the proposed method. The results show that our method is able to predict the generation and dissipation processes of evacuation demand, as well model how severely stations will be affected by given disruptions. One general observation we draw from the results is that the most vulnerable stations under disruption, where the locations peak evacuation demand occurs, are mainly turn-back stations, closed stations, and the transfer stations near closed stations. This paper provides new insight into evacuation demand prediction under disruptions. It could be used by transport authorities to better respond to the urban rail transit system disruption.