Journal of Global Optimization
IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews
JADE: adaptive differential evolution with optional external archive
IEEE Transactions on Evolutionary Computation
Adaptive particle swarm optimization
IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
A novel set-based particle swarm optimization method for discrete optimization problems
IEEE Transactions on Evolutionary Computation
Clustering-Based Adaptive Crossover and Mutation Probabilities for Genetic Algorithms
IEEE Transactions on Evolutionary Computation
| Hi-index | 0.00 |
Crossover is a very important operation in current differential evolution (DE) algorithms. The existing crossover strategies in DE show promising effects especially when the algorithms are applied to separable functions. However, the operation fails to work well when applied to the ill-conditioned and inseparable problems because the recombination of good genes is no longer promising for generating better individuals when the genes are highly correlated. Thus it is possible to use coordinate system rotation strategy which makes the variables be less correlated to improve the performance of the crossover operation. In this paper, we propose to use the principal component analysis (PCA) technique to rebuild a coordinate system. With this system, the correlations among variables are decreased for the crossover operation of DE. In every generation, the population and the mutated population are rotated into the new coordinate system to perform the crossover and then the newly generated population is rotated back to be evaluated. The PCA-based crossover is tested on the JADE algorithm. Experimental results show that the proposed method is quite promising on some ill-conditioned and inseparable functions.