Customized bus routing problem considering co-opetition with taxis at transport hubs

At transport hubs, long taxi queues greatly degrade the passenger transfer experience, particularly at night when conventional buses and subways stop running. To address the difficulties in hailing taxis at transport hubs and improve hub transfer efficiency, we propose a combined mode referred to as the Customized Bus Plus Taxi (CBPT) service. In this mode, the customized bus (CB) transports passengers from the hub to bus stops closer to their destinations, where hailing a taxi is easier, which reduces both passenger travel time and costs. A logit model is constructed to capture passenger choices between CBPT services and taxi-alone services at hubs. A Mixed-integer Non-linear Programming (MINLP) model is formulated to solve the joint optimization problem for CB route and frequency design, considering time-varying demand. A customized gradient algorithm and a rolling horizon scheme are developed to solve the MINLP model. The application to Shanghai Hongqiao Hub demonstrates that the algorithm can solve a real-world problem in reasonable time. The results indicate that the CBPT service mode can reduce total costs by 43 % and waiting time costs by 84 % compared to using taxi services alone. The average taxi queuing time at the hub decreases from 50 mins to 8 mins. The CBPT service mode can greatly improve passenger transfer efficiency at the hub.