Risk Analysis of Key Nodes of Sea Lanes Using Advanced Bayesian Network

The risk analysis of key nodes is the foundation and important link to ensure the safety of the sea lanes. To analyze the risk and provide a basis for safety management, this paper proposes a risk assessment model based on a multi-state fuzzy Bayesian network. Firstly, a fault tree model is built for risk events, revealing the causal relationships between the risk events and the influential factors. Secondly, the multiple fault state of nodes is described by fuzzy numbers. The multi-state fuzzy conditional probability table is applied to describe the uncertain logical relationship between nodes. An expert investigation method based on a confidence index is proposed to establish a multi-state, conditional probability interval table that describes the logical relationship between variables. Finally, the model parameters are determined by obtaining the exact values of the conditional probability interval table using the α-weighted valuation method for defuzzification. The risk probability distribution based on prior knowledge and known evidence is calculated. The key risk factors are then identified. Taking the risk analysis for key nodes of sea lanes as an example, the application results demonstrated the feasibility and efficiency of the proposed method.