基于贝叶斯网络的内河搁浅船舶应急决策方法

Emergency Decision-Making Method for Grounding Ships Handling in Inland Waterway Based on Bayesian Network

为使内河搁浅船舶快速有效地进行应急处置,提出一种基于贝叶斯网络的应急处置决策方法,该方法具有直观性强、能定量分析各种应急方案和能处理不确定性信息等优点。通过对江苏海事局2009—2012年的搁浅船舶应急处置数据进行统计分析,获得搁浅船舶的应急处置方案和相对应的影响因素值。提出搁浅船舶的三层应急决策框架,即影响因素层、决策准则层和决策方案层。在该决策框架的基础上建立搁浅船舶应急处置模型,并介绍各决策层的因素及其含义;同时,利用条件概率表建立相互间的定量联系。以一艘在长江内运营的典型搁浅船舶为例进行案例验证,结果表明,提出的贝叶斯决策模型能很好地处理搁浅船舶的应急处置。

In order to handle grounding accidents timely and effectively in inland waterway transportation,a Bayesian network based decision-making method is proposed,which owns the distinguishing characters of intuitive,quantitative assessment on alternatives and being able to deal with uncertainty. The data of emergency response to grounding ships from 2009 to 2012 in Jiangsu MSA are collected and processed with statistical analysis to find the portfolio of response measures and corresponding influencing factors. The three-layer emergency decision-making framework for grounding ships is proposed,which consists of the influencing factors layer,the decision attribute layer and the decision layer. The emergency decisionmaking model is established in the three-layer framework. The factors of each layer are explained in detail and the relationships between the influencing factors are established using conditional probability tables. A typical grounding accident in Yangtze River is used to verify the proposed model. The result demonstrates that the proposed Bayesian network model is useful for emergency response to grounding accidents.