Abstract: The computational speed in the feature selection of Mahalanobis-Taguchi system (MTS) using standardbinary particle swarm optimization (BPSO) is slow and it is easy to fall into the locally optimal solution. Thispaper proposes an MTS variable optimization method based on chaos quantum-behavior particle swarm. Inorder to avoid the influence of complex collinearity on the distance measurement results, the Gram-Schmidtorthogonalization method is first used to calculate the Mahalanobis distance (MD) value. Then, the optimalthreshold point of the system classification is determined through the receiver operating characteristic (ROC)curve; the misclassification rate and the selected variables are defined; the multi-objective mixed programmingmodel is built. The chaos quantum-behavior particle swarm optimization (CQPSO) algorithm is proposed to solvethe optimization combination, and the algorithm performs binary coding on the particle based on probability.Using the optimized combination of variables, a new Mahalanobis-Taguchi metric based prediction system isestablished to complete the task of precise discrimination. Finally, a fault diagnosis for the steel plate is taken asan example. The experimental results show that the proposed method can effectively enhance the iterative speedand optimization precision of the particles, and the prediction accuracy of the optimized MTS is significantlyimproved.

Key words: Mahalanobis-Taguchi system (MTS), variable selection, chaos quantum-behavior particle swarm,optimization