Abstract
A practical problem in the operations of Line 2 of the Shanghai Metro in China—the severe congestion at the Guanglan Road station—led to the proposal of a total cost (including both operator’s and users’ costs) minimization model for redesigning short-turn operations. Existing studies were extended mainly by relaxing the assumption that the full-length route must be operated because this assumption was determined to be needlessly strict and could, therefore, lead to inferior solutions. Flexible models were designed to accommodate more complex combinations of train routes and to calculate potential waiting costs if no direct services were available for some origin–destination pairs. The method was tested through real-world case studies, and major findings were interpreted. Future research directions are also suggested.