【摘要】 在多准则决策中,指标间完全独立是少见的,指标相关却是一种常态.指标对评价目标直接"贡献"和该指标通过其他指标对评价目标产生的间接"贡献"是确定关联指标权重的关键.本文利用经济学中的"弹性系数"对指标"贡献"进行测度.根据确定独立指标权重的方法可获得指标直接弹性系数;通过绘制指标间的因果关系图,构造指标弹性系数矩阵,通过矩阵运算可得到各指标的间接弹性系数.最后根据各指标总的弹性系数计算指标权重.结果显示,指标权重是该指标弹性系数占所有指标弹性系数总和的比例,独立指标权重是关联指标权重的特例.实例表明,在指标关联条件下,忽视指标关联得到的权重存在较大误差,具有重要影响作用的指标权重会被低估.更多还原

【Abstract】 In multi-criteria decision making,full independence is rare among indicators,but it is a norm that indicators are related.The key to determine the related indicators weights is the direct "contributions" that the indicators work on the target and the indirect "contributions" that the indicators work on the target through other indicators.In this paper,it uses economics "elasticity coefficient" to measure indicators "contributions".It can get index direct elasticity coefficient by using the way of determining independent index weights;It can construct index elasticity coefficient matrix by drawing a causal relationship diagram among indicators,and can get index elasticity coefficient among indicators through matrix operations.Finally,it calculates the index weights according to every indicator’s total elasticity coefficient.The results show that the index weight is the proportion of the index elasticity coefficient in the sum of index elasticity coefficient,the independent index weight is a special case of related index weight.The example shows that,under the condition of related indicators,to get index weght by ignoring related indicators will cause huge errors,the index weights that the indicator have an important influence will be underestimated.更多还原