多元判断偏好集结的混合群决策过程研究

摘要

Aiming at the different preference expressing in the evaluation process,a two-phase aggregation method about group decision-making is proposed:personal preference conversion and multiple experts’attribu-ting.The step of personal preference conversion aims to convert various experts’different preferences in differ-ent forms,such as fuzzy preference relations,utility function,partial selection,preference ordering,and lin-guistic preference relations to normalized numerical absolute preferences.And in the expressions of the expert weight of relative preferences,the fuzzy ordered weighted averaging (OWA )operator and the linguistic ordered weighted averaging (LOWA )operator are adopted.Multiple experts’attributing is arranged by the expert weight of arithmetic progression,and then the weight is dealt to normalization.According to the total score of the objects,the best one which has the highest score is chosen,therefore the democratic choice of the expert groups is gotten.Finally, apply the two step method into an example about a talents’selecting in a company.Compared with the previous me-thods in selection,the proposed method highlights the decision under the free,democratic preference.%针对评价过程中出现的不同偏好表达，提出了两阶段集结的群体决策方法：个人偏好转换和多专家集结。个人偏好转换把不同专家做出的模糊偏好关系、效用函数、部分选择、偏好排序、语言偏好关系等表达方式，转换为归一化的数值绝对偏好，并在专家相对偏好的权重表达中，采用模糊有序加权平均（ordered weighted averaging，OWA）算子和语言有序加权平均（linguistic ordered weighted averaging，LOWA）算子；多专家集结则根据等差数的专家权重排列，并将权重做归一化处理，得出总分后即从可行方案中选出最优的方案，分数越高表明专家群体越为偏好，从而获得多元偏好表达的混合专家群体的民主选择结果。将该两步集结的方法用于某公司的人才综合评估中。与以往传统的选择方法相比，突出了决策者偏好表达的自由、民主的特征。