Abstract: The main work of this paper is to determine the weights for Group Decision based on the weighting of the indices. Firstly, we get the reasonable range of the indices based on the weights of the expert. And then, this paper builds the optimization model to calculate the combination weight. At last, case studies testify the validity and feasibility of this method. The contribution characteristics lie on two aspects. Firstly, this paper makes the consistency test on the indicators take the place of the weight vector of the experts, and avoids the mistake of the effective information deleted and the amplification of invalid information. Secondly,this paper gets the combination weight based on the weighting of the indices,and solves the weight problem in group decision-making.

Key words: group decision, combination weight, indicates, optimization model

摘要： 本文的主要工作是从指标层入手对群决策中的赋权方法进行研究。首先,根据多位专家给出的某一指标的多个权重,通过一致性检验确定该指标的合理取值区间;然后,以组合权重与通过一致性检验的专家权重之间的偏差最小为目标函数建立优化模型,求解组合权重;最后,通过算例验证了模型的有效性和可行性。本文的主要创新与特色有两点:一是从指标层面对专家权重的一致性进行检验,与根据专家的权重向量的一致性检验相比,更加灵活,且避免了对有效信息的删除和无效信息的放大;二是从指标层面确定指标的组合权重,解决了群决策中的组合赋权问题,改变了专家权重分配对组合权重的影响问题。

关键词: 群决策, 组合赋权, 指标层, 优化模型