Research Of Polar Shipping Risk Assessment And Decision System Based On Bayesian Network

In recent years,as global warming has made parts of the Arctic ocean ice-free,commercial shipping has become predictable.The Arctic shipping route provides a convenient transportation route for trade between China and Europe and North America,and it is also an important supplementary section of "One Belt And One Road" in China.With the improvement of navigation conditions in the Arctic and the deepening of people’s exploration of the Arctic,the risk assessment technology of polar shipping has become more and more important.At present,the Arctic waters still lack the necessary navigational AIDS,and ships are prone to accidents under the influence of bad weather.In addition,most countries are not familiar with the navigation environment in the Arctic,and many sailors lack the experience of sailing in the Arctic,which makes it difficult to respond effectively to risks when they occur.Therefore,further study on the risk assessment technology of polar shipping will help to reduce the occurrence of polar shipping accidents.This paper firstly analyzes the advantages and disadvantages of the existing polar shipping risk assessment methods,and puts forward the bayesian network method for polar shipping risk assessment in view of the numerous factors affecting the Arctic navigation environment and the mutual influence among them.Based on the study of the learning algorithm of bayesian network,it is found that bayesian network can find the potential correlation of factors from the sample data,making the evaluation results more accurate and comprehensive.Then,the paper on how to choose the main factors affecting the safety of shipping is analyzed,using the Arctic historical accident frequency information of risk factors,based on grey correlation analysis of risk factors are compared,and the weights finalized with environment,vessels,personnel,accident four aspects on the basis of 20 major influencing factors.What’s more,aiming at the problem of direction edge missing in the learning results of bayesian network structure,the paper designs a structure establishment method integrating prior knowledge,which will explain the structure model and the combination of Structure Expectation Maximization(SEM)algorithm’s bayesian network structure to form a combined network structure.Through the analysis,it is found that SEM algorithm can find the potential correlation of factors from the data features,and the interpretation structure model can make up for the lack of directed edges generated by SEM algorithm,and the two algorithms complement each other.On the basis of the establishment of the combined network structure,the probability distribution of nodes was obtained by learning the parameters of the structure,and the node was deduced by using the join tree algorithm.Finally,the risk assessment and decision model were established,and the validity of the model was verified by MATLAB simulation.Finally,aiming at the problem of lack of effective decision Suggestions after risk assessment,the paper develops the design and implementation of polar shipping risk assessment and decision system based on the bidirectional inference characteristics of bayesian network.The system is developed based on QT and MATLAB platform,which can assess the risk of current shipping status and give corresponding auxiliary Suggestions.Combined with the actual situation of the Arctic ocean,the performance of the system was verified by the sample data test,and the results showed that the system could make a comprehensive and accurate assessment of the current shipping state,and give effective Suggestions,so as to provide technical support for ensuring shipping safety and preventing accidents in the polar region.