Simulation Experiments for Improving the Consistency Ratio of Reciprocal Matrices
Daji Ergu, Gang Kou, Yi Peng, Xinfeng Yang · International Journal of Computers Communications & Control · 2014
The consistency issue is one of the hot research topics in the analytichierarchy process (AHP) and analytic network process (ANP). To identify the mostinconsistent elements for improving the consistency ratio of a reciprocal pairwisecomparison matrix (PCM), a bias matrix can be induced to efficiently identify themost inconsistent elements, which is only based on the original PCM. The goal of thispaper is to conduct simulation experiments by randomly generating millions numbersof reciprocal matrices with different orders in order to validate the effectiveness ofthe induced bias matrix model. The experimental results show that the consistencyratios of most of the random inconsistent matrices can be improved by the inducedbias matrix model, few random inconsistent matrices with high orders failed theconsistency adjustment.