With the increase in urbanization, transit networks are presenting characteristics, such as complex configurations, high station plan density, and high ridership level. Most existing studies focus on designing a new transit network in a planning area. By contrast, research on adjusting or redesigning an existing network to improve its performance remains insufficient. To fill this gap, this study proposes a methodological framework that dynamically optimizes an existing transit network, considering frequency and structure simultaneously according to ridership variation. The proposed framework is divided into frequency and structure optimization. First, a criterion is introduced to measure the necessity of network optimization. Second, the particle swarm algorithm is adopted to optimize frequency. Third, multiple strategies, including deleting, extending, and adding lines, are proposed to optimize the existing transit network. These dynamic optimization procedures are implemented with multicycle iterations. The transit network of the Beijing GuoMao–TongZhou commuting corridor is studied on the basis of transit smart card data. Results show that the proposed methodological framework is effective in transit network optimization.

CICTP