Integrating Naive Bayesian network and evidential reasoning for risk assessment in general cargo ship pre–stowage planning

In the current General Cargo Ship Pre-Stowage Planning (GCSPP), efficiency and economy are the two main dominant influential factors. Given the increased operational uncertainty and safety priority, a new solution that can incorporate the safety factor into the planning becomes highly demanded with urgency. Such a new solution is critical for reinforcing compliance with safety standards and regulations, ultimately enhancing the competitiveness of the general cargo shipping industry. To bridge this gap, we have developed an innovative data-driven quantitative GCSPP assessment framework that combines Naïve Bayesian network (NBN) modeling with evidence reasoning (ER) techniques. This framework focuses on quantitative analysis, supplemented by standardized qualitative input conversion, to incorporate safety, economy, and efficiency into the evaluation and selection of GCSPPs. The framework incorporates dynamic risks associated with ships navigating diverse environments into the planning process. Fourteen Risk influencing Factors (RIFs) related to GCSPPs were identified from general cargo ship accident reports from 2007 to 2021 and incorporated into a NBN model. Using forthcoming voyage information, navigation scenarios are constructed to determine the safety requirements for pre-stowage planning. These requirements are translated into weighting factors and, together with efficiency and economy, integrated into the ER method to support multi-criteria decision-making for GCSPPs. This enables the identification of the optimal pre-stowage plan for a specific voyage by jointly considering safety, efficiency, and economy. The findings show that, for the same ship carrying the same cargo, the most suitable pre-stowage plan varies with voyage-specific safety requirements. The study further contributes a robust solution to complex decision-making problems involving multiple dynamic attributes through the integrated application of NBN and ER techniques.