摘要
针对区间型多属性群决策中的专家赋权问题,提出了基于矩阵相似度的专家客观权重的确定方法。首先,通过对加权规范化决策矩阵进行两两比较得到相似度矩阵;然后对相似度矩阵每行或每列求和,计算出粗权重,并且给出了区间型理想矩阵的概念;最后,通过计算加权规范化决策矩阵与区间型理想矩阵的投影值得到专家权重。与区间型理想解相比,文中给出的区间型理想矩阵考虑了专家的决策权力,更加准确。算例验证了该方法的有效性。
A determination method of expert objective weight based on matrix similarity is proposed for interval multi-attribute decision-making. Firstly, similarity matrices are obtained by comparing weighted standardization de- cision matrices; secondly, every row or column of the matrix similarity is added to calculate the rough value of the weight of experts, noted as rough weight and then a concept of interval ideal matrix is also presented; finally, by comparing the projection of weighted standardization decision matrix and the interval ideal matrix, the weight of ex- perts is obtained. Compared with interval idea solution, the interval ideal matrix in this paper is more accurate. Ex- amples demonstrate the superiority of the proposed method.
出处
《电子科技》
2014年第2期7-9,共3页
Electronic Science and Technology
关键词
矩阵相似度
专家粗权重
区间型正理想矩阵
投影
matrix similarity
the estimated weights of decision makers
ideal decision matrix
projection
