对于层次分析法中不满足一致性要求的判断矩阵,提出了一种新的生成算法.此算法首先对目标因素集进行预排序生成初始判断矩阵,由上取整函数生成比较矩阵,然后由初始判断矩阵和比较矩阵构造相对误差矩阵,并在相对误差矩阵和矩阵相异度的控制下逐步调整初始判断矩阵,最后生成满足一致性要求和矩阵相异度最小的目标判断矩阵.仿真结果证实了所提出方法的可行性和有效性.
An algorithm is proposed for the non-consistent judgment matrix in AHP. A primary judgment matrix is generated firstly through pre-ordering the targeted factor set, and a compared matrix is built through the top integral function. Then a relative error matrix is created by comparing the compared matrix with the primary judgment matrix which is regulated under the control of the relative error matrix and the dissimilar degree of the matrix step by step. Finally, the targeted judgment matrix is generated to satisfy the requirement of consistence and the least dissimilar degree. The feasibility and validity of the proposed method are verified by simulation results.