中文摘要：

运用马氏距离替代欧式距离改进传统的TOPSIS方法，解决当属性间存在线性相关时欧式距离失效的缺陷；充分考虑对立集合并引入联系向量距离，解决可能存在的方案距离正理想解和负理想解距离都近的缺陷．然后通过决策者偏好系数将马氏距离和联系向量距离所得结果合成新的相对贴近度，从而同时克服传统TOPSIS方法的以上两个缺陷．最后通过供应商选择的实例来验证方法的有效性．

英文摘要：

To solve the defects of Euclidean distance failure When there is a linear correlation between attributes, the method of Mahalanobis Distance instead of Euclidean distance is adopted to improve the traditional TOPSIS in this paper; And it fully considers the Collection of Opposites and introduces the Connection Vector Distance to solve the possible defect that the distance between solution and positive ideal solution is close, and the distance between solution and negative ideal solution is also close. Then it combines the results obtained from Mahalanobis Distance and Connection Vector Distance by Decision-making Preference Factor into a new kind of relative similarity degree, which can overcome the two defects of the traditional TOPSIS method. Finally, it verifies the effectiveness of the method by the examples of the Supplier Selection.