中文摘要：

判断矩阵排序权重计算及其一致性检验可以归结为一个使一致性指标最小化的非线性、变量耦合的优化问题。提出一个利用基于序贯重点采样粒子滤波和Cholesky分解的分布估计算法计算排序权重的算法，该算法采用的概率模型是多峰的并考虑了变量之间相关性。文中对判断矩阵排序权重计算及一致性检验、基于序贯重点采样粒子滤波和Cholesky分解的分布估计算法进行了介绍，描述了排序权重计算及一致性检验算法，最后给出实验数据及其分析。实验结果验证了算法的有效性，并具有很高的精度和稳定性。

英文摘要：

The rank weights computing and the consistency check of judgement matrix can be attributed to a nonlinear, coupled variable optimization problem to minimum the consistency index.An algorithm for computing rank weights using estimation of distribution algorithm based on sequential importance sampling particle filter and Cholesky decomposition（PFEDA2） is presented in this paper.The probability model of this algorithm is multi-peak and the correlation between variables is considered.The method of rank weights computing, the consistency check and PFEDA2 is introduced, and then the algorithms of rank weights computing and consistency check are described.Finally, the experimental data and the analysis is given.Experimental results verify the effectiveness of the algorithm,and show this algorithm is high accuracy and stability.