Identifying Spatial–Temporal Characteristics and Significant Factors of Bus Bunching Based on an eGA and DT Model

Bus bunching is a common phenomenon caused by irregular bus headway, which increases the passenger waiting time, makes capacity uneven, and severely reduces reliability of service. This paper clarified process formation, analyzed variation characteristics in single day, different types periods, at stops, then concluded twelve potential factors. A hybrid model integrating genetic algorithm with elitist preservation strategy (eGA) decision tree (DT) was proposed. The eGA part constructs framework transforms factor identification into problem selecting fittest individual from population, while DT evaluates fitness. Model verification comparison were conducted based on real automatic vehicle location (AVL) data Shenzhen, China. results showed that proposed eGA–DT outperformed other frequently used extra (ET) models least 20% reduction MAE under routes, stops. Six factors, including sequence stop, headway dwell time previous travel between etc., identified to have significant effect bunching, great value for feature selection improve accuracy efficiency prediction real-time dispatching.