基于动态贝叶斯网络的北极东北航道关键海区通航可行性研究

A FEASIBILITY STUDY ON NAVIGATION IN A KEY SEA AREA OF THE NORTHEAST PASSAGE OF THE ARCTIC BASED ON THE DYNAMIC BAYESIAN NETWORK

全球的持续变暖导致极地海冰不断融化,北极航道即将全面开通。针对北极东北航道关键海区通航的不确定性、数据的时序性等问题,提出基于动态贝叶斯网络的北极东北航道关键海区通航可行性评估与预测技术。通过对航道通航关键海区的确定、指标的选取、贝叶斯网络的构建与参数学习、基于证据的推理计算、动态贝叶斯技术的检验等流程,建立了基于动态贝叶斯网络的风险评估与预测模型。结果表明,动态贝叶斯网络技术不仅能够很好地处理动态时序信息,且其在推理过程中结合了历史信息,具备信息的累积能力,降低了误差数据对最终结果的影响度,使结果更接近真实值,且动态贝叶斯网络技术的预测结果在短期内准确性较高,对判断北极东北航道的可通航性具有重要的参考意义。

Continued climate change has led to the melting of polar ice and the opening of the Arctic passage.Because of the uncertainty of navigability in a key sea through Arctic Northeast Passage and the typical time series nature of the statistical data, the assessment and forecasting technology for the risk affecting the navigability of the Northeast Passage based on Dynamic Bayesian Network is proposed. Through determination of the key sea and index;construction of the structure and parameter learning of the Bayesian Network;reasoning and computation based on evidence;and testing of the superiority of Dynamic Bayesian Network, a risk assessment and forecasting model based on Dynamic Bayesian Network is established. The results show that the Dynamic Bayesian Network technology can not only deal with the dynamic data but can also reduce the effect of inaccurate data on the results. The Dynamic Bayesian Network is superior to the traditional evaluation technology, and the predicted results of the model are more accurate in the short term.This result has an important significance for judging the navigability of the Arctic Northeast Passage.