Big data-driven booking consolidation and scheduling of launching service in Singapore Port

• Published in: 2024 IEEE Conference on Artificial Intelligence (CAI) pp. 272 - 277
• Main Authors: Lin, Yun Hui, Chua, Ping Chong, Yin, Xiao Feng, Wang, Zizhe, Li, Ning, Xiao, Zhe, Fu, Xiuju, Qin, Zheng
• Format: Conference Proceeding
• Language: English
• Published: IEEE 25.06.2024
• Subjects: big-data analytic; Boats; Delays; demand batching; launch boat operations; Real-time systems; Scheduling algorithms; Seaports; Task analysis; task scheduling; Trajectory; trajectory prediction
• DOI: 10.1109/CAI59869.2024.00060

Launching services provided by launch boat (LB) operators are indispensable for vessels in port areas. In the current business practice, the operator follows a "one-trip-per-booking" method, where each booking corresponds to a LB transporting passengers to their destination. Undoubtedly, this method does not efficiently utilize the LB's capacity. A more appealing approach is to consolidate or batch multiple service bookings into a single LB, enabling it to travel to multiple destinations within one trip. Using large-scale GPS data of LBs, we conduct data-driven analysis to gain insights into LB trajectory and traveling pattern. Based on them, we propose a real-time batching algorithm to consolidate a maximum of two bookings into a task with marginal service delay. We then address the scheduling of LBs to fulfill the consolidated tasks using rule-based real-time approaches. To validate our proposed framework, we conduct a case study in Singapore Port. The results show that after implementing the data-driven batching and scheduling algorithms, we achieve a reduction of more than 25% in the traveling distances of LBs, while maintaining a high level of service quality for passengers.