基于贝叶斯网络的船舶港口国监督检查滞留风险研究

针对港口国监督(Port State Control, PSC)检查的复杂性和不确定性, 基于贝叶斯网络理论构建船舶PSC检查滞留风险分析模型. 以东京备忘录(Tokyo MOU)中2014~2017年船舶PSC检查样本数据为基础, 运用R语言bnlearn包进行贝叶斯网络的结构及参数学习. 同时分别执行贝叶斯网络的正向、逆向推理, 定量表示各风险因素与滞留结果之间的相互作用关系, 找出导致船舶滞留的高风险因素, 实现不确定环境下船舶PSC检查滞留风险的全面动态分析. 实证表明, 模型具有较高的精确度, 可为检查人员的滞留决策及航运公司的安全风险管理提供有效依据.

Detain risk of ship’s port state control inspection based on Bayesian networks

In view of the complexity and uncertainty of port state control (PSC) inspection, in this paper a dynamic analysis model is constructed for the detain risk of ship’s PSC inspection using the approach of Bayesian networks. Upon the sample data of ship’s PSC inspection in Tokyo MOU from 2014 to 2017, the structure and parameters of Bayesian networks are studied using the bnlearn package of R language. Mean- while, the forward and backward reasoning of Bayesian networks are carried out to quantitatively express the interaction among the risk factors and detention, and the high-risk factors that affect ship’s detention are determined. A comprehensive and dynamic analysis of the detain risk of ship’s PSC inspection under uncertain environment is conducted. The experimental results show that the model has high accuracy and can provide a valid basis for inspectors’ detention decision-making as well as safety risk management of shipping companies.