中文摘要：

文献检索结果表明，在综合评价中，对权重的研究主要局限于新方法的开拓和优化方面，对其本身误差的研究较少涉及。鉴于权重的重要性，本文以序关系法为例，以蒙特卡罗分析法为工具，分别讨论了“噪声”对单个指标权重和全部指标权重的影响。分析过程显示，每一次判断所产生的“噪声”都会沿着“指标链”进行累积，并极有可能会产生“蝴蝶效应”，最终导致评价结果的颠覆。而合理的减少指标个数和有效的提高指标重要性之比的赋值（尤其是最初几次的赋值的准确性）是获得相对准确指标权重的关键。

英文摘要：

It can be demonstrated from the research data that, in the field of comprehensive evaluation, the research on the weight is confined to the exploration and optimization of new methods, while the research on its own error is less involved. Considering the importance of the weight, this paper focus on rank correlation method and use Monte Carlo analysis method as a tool to conduct quantitative analysis on the error of single index weight and overall indexes weight respectively. The analysis shows that the ＂noise＂ arising from every judgment will be accumulated along the ＂index chain＂ , and the ＂noise＂ may produce ＂the butterfly effect＂ , which will lead to the subversion of the evaluation results. Furthermore, the key to determine the weight by effectively using rank correlation method are the reasonable reduction of the number of index and effectively improve the accuracy of the judgment of the relative importance （especially the initial several times）.