Modelling node-selection behaviour in subway stations.

Node-selection behaviour in subway stations directly impacts the utilisation and capacity of equipment and facilities. Although this topic has been extensively studied in various fields, existing studies are limited to environment characteristics while little research has been undertaken from the perspectives of pedestrian and interaction characteristics. In this study, pedestrian node-selection behaviour at automatic fare gates (AFGs) in subway stations was investigated and a multi-nomial logit model was developed that incorporates environment, pedestrian and interaction characteristics. For model estimation and validation, a detailed site survey was carried out to collect data on pedestrians' actual AFG selection at three Beijing subway stations. The prediction accuracy of the model reached 90·90%. Application of the model is demonstrated by detailed analyses of influence factors from four aspects along with the utilisation optimisation of AFGs in Beijing West subway station. [ABSTRACT FROM AUTHOR]