TY - GEN
T1 - Towards a genetic direct least squares prioritization method for pairwise reciprocal matrices
T2 - 2016 International Conference on Machine Learning and Cybernetics, ICMLC 2016
AU - Yuen, Kevin Kam Fung
N1 - Publisher Copyright:
© 2016 IEEE.
PY - 2016/7/2
Y1 - 2016/7/2
N2 - Analytic Hierarchy Process is a popular decision tool that ranks the aggregated results from prioritizing the pairwise reciprocal matrices judged by the users. Saaty's Eigen method for prioritizing pairwise reciprocal matrix leads to the possibility of rank reversal, and is still one of the unsettled issues, although many applications of this method have been made. This study proposes the genetic direct least squares prioritization (GDLSP) method to solve the pairwise reciprocal matrices. The direct least squares prioritization (GDLSP) method produces less total square variance between ground pairwise reciprocal matrix and its prioritization results. Genetic algorithm is a promising method to effectively solve the direct least squares problem with proposed design to handle the DLS constraint. A financial investment problem regarding asset portfolio selection using Saaty's Eigen method is revised by the proposed method.
AB - Analytic Hierarchy Process is a popular decision tool that ranks the aggregated results from prioritizing the pairwise reciprocal matrices judged by the users. Saaty's Eigen method for prioritizing pairwise reciprocal matrix leads to the possibility of rank reversal, and is still one of the unsettled issues, although many applications of this method have been made. This study proposes the genetic direct least squares prioritization (GDLSP) method to solve the pairwise reciprocal matrices. The direct least squares prioritization (GDLSP) method produces less total square variance between ground pairwise reciprocal matrix and its prioritization results. Genetic algorithm is a promising method to effectively solve the direct least squares problem with proposed design to handle the DLS constraint. A financial investment problem regarding asset portfolio selection using Saaty's Eigen method is revised by the proposed method.
KW - Analytic Hierarchy Process
KW - Asset Allocation
KW - Constrained Genetic Algorithm
KW - Pairwise comparison
KW - Prioritization
UR - http://www.scopus.com/inward/record.url?scp=85021061135&partnerID=8YFLogxK
U2 - 10.1109/ICMLC.2016.7860913
DO - 10.1109/ICMLC.2016.7860913
M3 - Conference Proceeding
AN - SCOPUS:85021061135
T3 - Proceedings - International Conference on Machine Learning and Cybernetics
SP - 275
EP - 279
BT - Proceedings of 2016 International Conference on Machine Learning and Cybernetics, ICMLC 2016
PB - IEEE Computer Society
Y2 - 10 July 2016 through 13 July 2016
ER -