Abstract: To the problem of multiple methods without unified comparison standards in comprehensive evaluation, we choose four evaluation methods highlighting the difference among objects to analyze their stability and difference highlighting capability using the approach of stochastic simulation. The conclusions are drawn that the stability of four methods from high to low is variance, maximum deviation, entropy, deviation maximization method, and the more stable of a method, the less capability of highlighting difference. This research not only can test and verify the features of four methods which could provide reference for evaluators of methods selecting, but also can provide technology support, that is stochastic simulation method, for other evaluation methods comparison.

Key words: comprehensive evaluation, difference driving, stochastic simulation, stability, difference highlighting

摘要： 针对综合评价中方法种类繁多且无统一比较标准的问题,选取了四种突出被评价对象之间差异的评价方法,采用随机模拟的方式分别从评价方法的稳定性及对差异的凸显能力两个方面进行了比较分析。得出了四种方法稳定性由高到低分别为均方差法、最大离差法、熵值法、拉开档次法,且评价方法的稳定性越高,则其对差异的凸显能力反而越差的结论。该研究不仅验证了差异驱动型评价方法的相关特性,为评价者关于评价方法的选取提供了参考意见,而且随机模拟方法的应用,可为类似的多评价方法的比较问题提供技术参考。

关键词: 综合评价, 差异驱动, 随机模拟, 稳定性, 差异凸显