关键词: 交通运输工程；海运通道；贝叶斯网络；因素识别；主成分分析；韧性

Abstract: The safety and unimpeded passage of sea lanes is the core guarantee for the stable development of the maritime transport industry. In order to evaluate the resilience level of sea lanes scientifically, a resilience evaluation framework based on principal component analysis, fuzzy logic and Bayesian network model was proposed. Firstly, by analyzing the relevant accident reports and literature, the key factors affecting the resilience level of sea lanes was identified on the basis of the dually driven perspective. Then, in order to make up for the defects of the correlation between the influencing factors and fully mine the sample information, principal component analysis was used to screen the initial indexes. Finally, the index weights obtained by principal component analysis were used as the quantitative input of the interrelationships between influencing factors in Bayesian networks, thus taking into account the importance of influencing factors and effectively overcoming the defect of ignoring the importance of indexes in traditional Bayesian networks. The research results show that the proposed resilience evaluation framework can measure the resilience level of sea lanes more effectively than other evaluation methods do, which provides an important theoretical basis and decision-making reference for the resilience evaluation of sea lanes.

Key words: traffic and transportation engineering; sea lanes; Bayesian network; factor identification; principal component analysis; resilience