AI for PSC at Hong Kong
This project aims to predict the risk of the visiting foreign-flag ships at Hong Kong, evaluated by the number of deficiencies and detention probability in PSC inspection during this visit. More specifically, ships currently berthing at the Port of Hong Kong are constantly downloaded from the website of the Marine Department (MD) of HKSAR, China. Then, more specific ship features, including ship generic features such as ship type, ship dimension, keel laid date, total previous detentions, casualties involved, and flag change times, as well as historical PSC inspection information regarding last inspection time and condition and the ship’s company/RO/flag performance, are searched from relevant databases such as the online public PSC database provided by the Tokyo MoU. Based on the features, ship deficiency number and detention probability in the inspection during this visit are predicted using state-of-the-art artificial intelligence (AI) model. The normalized predicted deficiency number and detention probability are combined in a weighted sum manner (with 60% of the weights given to the normalized predicted deficiency number and 40% of the weights given to the normalized predicted detention probability), which is the final predicted ship risk. Finally, the predicted risk of the ships currently at the Port of Hong Kong, together with their main features, is output to an Excel file. The file includes the ships’ identity, type, flag, name of agent, last port of call, current location, ship risk profile, inspection priority, inspection time window, and the detailed prediction results given by the AI models.
This project is sponsored by the Policy Innovation and Co-ordination Office (PICO) of the Government of the HKSAR (Project number: 2020.A6.148.20A). The research team is led by Professor Shuaian (Hans) Wang from the Department of Logistics and Maritime Studies at The Hong Kong Polytechnic University. The team members are:
Jiannong Cao, Member of Academia Europaea, The Hong Kong Polytechnic University, HKSAR, China
Kjetil Fagerholt, Professor, Norwegian University of Science and Technology, Norway
Gilbert Laporte, Foreign Member of the US Academy of Engineering, University of Bath, UK
Xi Luo, Research Scholar, The Hong Kong Polytechnic University, HKSAR, China
Haoyu (Mok) Mo, Research Assistant, The Hong Kong Polytechnic University, HKSAR, China
Chuansheng Peng, Chief Technology Officer, China Waterborne Transport Research Institute, Ministry of Transport, China
Xiaobo Qu, Member of Academia Europaea, Chalmers University of Technology, Sweden
Defeng Sun, Chair Professor, The Hong Kong Polytechnic University, HKSAR, China
Xuecheng Tian, Research Scholar, The Hong Kong Polytechnic University, HKSAR, China
Ran (Angel) Yan, Research Scholar, The Hong Kong Polytechnic University, HKSAR, China
Dong Yang, Assistant Professor, The Hong Kong Polytechnic University, HKSAR, China
Lu Zhen, Professor and Dean of School of Management, Shanghai University, China
Moreover, Captain Z. Ni, Mr. Chiu-sum Mok, Senior Surveyor Mr Z.H. Huang, and Senior Surveyor Mr Meng Gong from Hong Kong Marine Department, Dr. Stephen Li from The Hong Kong Polytechnic University (former General Manager at Hong Kong Marine Department), and Captain Jian Feng Zhou, Captain Chang Zheng Chen, Captain Hongpeng Liu, Captain Zhang, and Mr Tim T.Y. Wang from Wah Kwong provided valuable feedback on this project.
This project improves the current Tokyo MoU practice by identifying 25% more deficiencies and 50% more detentions. This project was developed to improve the PSC efficiency at the Port of Hong Kong and can also be applied by other PSC authorities in the Tokyo MoU region. In case of enquiries or feedback, please contact Professor Shuaian (Hans) Wang at hans.wang@polyu.edu.hk