TY - GEN
T1 - Research on Dynamic Bus Station Location Selection for Demand-Responsive Transit Based on DBSCAN and K-Means Clustering Algorithms
AU - Lu, Di
AU - Zhang, Xiqiao
AU - Lin, Jiadong
AU - Li, Yetong
N1 - Publisher Copyright:
© 2026 IEEE.
PY - 2026
Y1 - 2026
N2 - Addressing the contradiction between fixed conventional bus stations and dynamic passenger flow distribution, as well as the high operational costs and scalability challenges of existing demand-responsive transit systems, this paper proposes a hybrid clustering algorithm combining DBSCAN and K-means for dynamic station location selection. The algorithm first filters out demand points within a 500 m service radius of fixed stations. After DBSCAN removes noise points, K-means iterative clustering identifies optimal centers. Station locations are then adjusted based on actual road networks. Using Harbin's ride-hailing peak-hour data as a case study, the algorithm ultimately identified 11 dynamic stations. This approach achieves a balance between service coverage and operational efficiency, providing effective support for constructing a hierarchical public transportation service network.
AB - Addressing the contradiction between fixed conventional bus stations and dynamic passenger flow distribution, as well as the high operational costs and scalability challenges of existing demand-responsive transit systems, this paper proposes a hybrid clustering algorithm combining DBSCAN and K-means for dynamic station location selection. The algorithm first filters out demand points within a 500 m service radius of fixed stations. After DBSCAN removes noise points, K-means iterative clustering identifies optimal centers. Station locations are then adjusted based on actual road networks. Using Harbin's ride-hailing peak-hour data as a case study, the algorithm ultimately identified 11 dynamic stations. This approach achieves a balance between service coverage and operational efficiency, providing effective support for constructing a hierarchical public transportation service network.
KW - DBSCAN Algorithm
KW - Demand Response Transit (DRT)
KW - Dynamic Station Location Selection
KW - Hybrid Clustering Algorithm
KW - K-means Algorithm
UR - https://www.scopus.com/pages/publications/105044892178
U2 - 10.1109/RAITS68656.2026.11580296
DO - 10.1109/RAITS68656.2026.11580296
M3 - 会议稿件
AN - SCOPUS:105044892178
T3 - Proceedings - 2026 International Conference on Robotics, Automation and Intelligent Transportation Systems, RAITS 2026
BT - Proceedings - 2026 International Conference on Robotics, Automation and Intelligent Transportation Systems, RAITS 2026
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2026 International Conference on Robotics, Automation and Intelligent Transportation Systems, RAITS 2026
Y2 - 23 January 2026 through 25 January 2026
ER -