An Improved Analytic Hierarchy Process

S. Su, J. Yu, and J. Zhang (PRC)


Decision analysis, analytic hierarchy process, judgment matrix, consistency, and accelerating genetic algorithms


The traditional AHP requires decision makers or experts remain consistent in making pairwise comparisons. Accurate expression of relative preferences on criteria is difficult due to the limitations of the 9-value scale of Saaty. Thus, the judgment matrix has definite subjectivity and inconsistency and its consistency is usually not satisfying. The purpose of this paper is to establish a scientific and suitable method for overcoming the weaknesses of AHP. To do this, this paper proposes a newly improved AHP (i.e. AGA-EAHP). It gains judgment matrix gangs using the extended-value scale replacing the 9-value scale firstly. And then it searches for the best consistency judgment matrices from the judgment matrix gangs with accelerating genetic algorithms and works out the weights of various criteria. Empirical analysis of AGA-EAHP and other methods shows that AGA-EAHP is better. It always can find the optimal consistency judgment matrix.

Important Links:

Go Back