Stochastic optimization model for ship inspection planning under uncertainty in maritime transportation

Maritime transportation plays a significant role in international trade and global supply chains. Ship navigation safety is the foundation of operating maritime business smoothly. Recently, more and more attention has been paid to marine environmental protection. To enhance maritime safety and reduce pollution in the marine environment, various regulations and conventions are proposed by international organizations and local governments. One of the most efficient ways of ensuring that the related requirements are complied with by ships is ship inspection by port state control (PSC). In the procedure of ship inspection, a critical issue for the port state is how to select ships of higher risk for inspection and how to optimally allocate the limited inspection resources to these ships. In this study, we adopt prediction and optimization approaches to address the above issues. We first predict the number of ship deficiencies based on a k nearest neighbor (kNN) model. Then, we propose three optimization models which aim for a trade-off between the reward for detected deficiencies and the human resource cost of ship inspection. Specifically, we first follow the predict-then-optimize framework and develop a deterministic optimization model. We also establish two stochastic optimization models where the distribution of ship deficiency number is estimated by the predictive prescription method and the global prescriptive analysis method, respectively. Furthermore, we conduct a case study using inspection data at the Hong Kong port to compare the performances of the three optimization models, from which we conclude that the predictive prescription model is more efficient and effective for this problem.