Combinatorial Optimization for the Guangzhou,China,Bus Rapid Transit System

Bus rapid transit (BRT) has become an efficient mode of public transportation to mitigate urban traffic congestion. Research on improving the performance of BRT systems is still in the early stages, despite booming growth of BRT projects. In some busy BRT systems, the challenge is to assign multiple substops reasonably and accurately to bus routes at the station. Bus queuing at the station is the major cause of traffic congestion in the Guangzhou, China, BRT (GBRT) system, which is the highest volume BRT system in Asia. To reduce queuing at the station and enhance the mobility of BRT buses, this study proposed a combinatorial optimization model (BCOM) to assign multiple substops to bus routes at the same station. The objective of the proposed method was to minimize the probability of bus queuing at BRT stations. The study used a novel genetic algorithm to obtain suboptimal solutions of the BCOM. Operating data from the GBRT system were analyzed, and it was found that the BRT bus arrival rate and dwell time obeyed a Poisson distribution and negative exponential distribution, respectively. A simulation tool for the GBRT system, based on VISSIM, was developed to evaluate the proposed solution. The simulation and application results demonstrated that the proposed solution could decrease bus queuing length effectively and could reduce bus dwell time and travel time and the service saturation of substops simultaneously. This study provides a useful and practical solution to improve the performance and operation of BRT systems and other traditional bus systems.