基于AHP—ML Ⅱ融合算法的先验分布研究

在缺乏可靠性试验数据的情况下,为得到可信的评定结论减少主观影响,需要从信息融合的角度出发,充分利用产品先验信息。以获得关于产品可靠性的更完整和更准确的判断信息,从而进一步形成对产品性能的可靠估计或预测。提出一种基于AHP-ML II的融合算法,对先验信息进行层次分析法处理后,再使用第二类极大似然估计法进行二次处理,通过不同的信息在融合后的先验分布中所起作用的大小,确定出各部分的权重因子,最后通过实例仿真分析得出该算法确实比融合前具有更好的合理性和可靠性。

Research on Fusion Method for Prior Distribution Based on AHP-ML II

In some circumstance,the analysis and evaluation should be based on the full use of information provided by all kinds of tests. The prior information must be reliable to execute fewer tests and get a reasonable prior distribution. It can take full advantage of the multi-source reliability information which be acquired at different time or from different place and get a more precise estimation of the product's reliability. For the lack of reliable test information,prior information should be fused and made full use of to decrease the subjective impact. Then a reasonable evaluation or prediction is gained and the information to make judgment is more acute. In this paper,a fusion algorithm of ML-II based on AHP is put forward. The prior information is processed by AHP and then by ML-II. The weight of different distributions is gained according to how distributions work after fusion. The result of simulation indicates that the method can make prior information before fusion more reliable and reasonable.