海盗袭击风险评估与反海盗应对策略评价

Pirate attack risk assessment and anti-piracy response strategy evaluation

为合理评估海盗袭击风险,实现反海盗应对策略科学评价,提出一种基于树形增强朴素(tree augmented native,TAN)贝叶斯网络的海盗袭击风险评估模型。根据西非地区海盗袭击事件的实际数据确定贝叶斯网络中各节点之间的概率依赖关系,保障评估结果的准确性。运用情景构建、灵敏度分析和Shapley值模型识别影响海盗袭击的关键因素,并科学评价反海盗应对策略的效果。结果表明:船速、攻击、海军护航是影响海盗袭击的关键因素,船员抵抗和海军护航是较为有效的反海盗应对策略。海盗袭击风险与采取反海盗应对策略之间存在着较为明显的线性影响关系,不同反海盗应对策略组合对海盗袭击的风险概率呈现差异性。该结果可为航运企业科学评估海盗袭击风险、制定反海盗策略提供参考依据。

With the aim of providing a reasonable assessment of pirate attack risk and realizing a scientific evaluation of anti-piracy response strategy,a tree augmented native(TAN) Bayesian network-based pirate attack risk assessment model is proposed.The actual data of pirate attacks in West Africa is used to determine the probability dependency relationship among nodes,which ensures the accuracy of the assessment results.The key factors influencing pirate attacks are identified using scenario construction,sensitivity analysis,and the Shapley value model,and the effectiveness of anti-piracy response strategy is scientifically evaluated.The findings indicate that ship speed,attack and naval escort are the key factors influencing pirate attacks,and crew resistance and naval escort are more effective anti-piracy response strategies.A significant linear relationship is found between the risk of pirate attacks and adopting anti-piracy response strategies,and the combination of different anti-piracy response strategies has different risk probability of pirate attacks.These results provide reference for shipping companies to scientifically assess the risk of pirate attacks and develop anti-piracy strategies.