A Multi-objective Evolutionary Algorithm of Principal Curve Model Based on Clustering Analysis

According to the traditional GA and EDA weakness, on the basis of MMEA, the orthogonal design initialization, convergence criterion and K-means clustering analysis method were introduced in this paper and it proposed a new model multi-objective evolutionary algorithm OMEA. The practice results showed that the OMEA had been greatly improved on both convergence and diversity of the solutions, reaching a good balance on diversity and convergence. Its comprehensive performance was better than the SPEA2, NSGA-II and other traditional multi-objective evolutionary algorithm.