灰色加权马氏链组合在航空装备事故预测中的应用

摘要

为提高传统灰色马氏链组合模型的航空装备事故预测水平，引入加权马氏链，提出了一种改进的灰色马氏链组合预测方法。该方法先通过建立灰色模型，提取事故序列的趋势信息，然后再利用灰色残余信息构建加权马氏链模型，合理发挥各步长马氏链的作用，以期准确刻画随机波动规律。为验证其有效性，在美国空军A级飞行事故万时率实际数据基础上，建立了灰色加权马氏链组合预测模型，结果表明，模型对2000年~2002年预测的相对误差平均控制在4.59%以内，远高于灰色叠加马氏链模型，所建的模型能够比较客观地反映航空安全的未来发展现状。%To improve the prediction level of traditional Grey Markov chains model in aviation accident,an improved Grey Markov chains combination prediction method based on Weighted Markov chains is proposed. In the method,Grey model is firstly established to extract the trend information from accident data series,and then weighted Markov chains model for the residual information of Grey model is constructed to reasonably play the role of various step-size Markov chains for describing the stochastic fluctuation law. To prove its validity,on the basis of a class flight accident 10,000-hour-rate actual data of USAF,Grey weight Markov chains combination model is established. The result shows that the average relative error of prediction of model during 2000 to 2002 is 4.59%,well above that of Grey stacked Markov chains model. The built model can objectively reflect the future development tendency of aviation safety.