Eigenvector Crossover in jDE100 Algorithm
Petr Bujok, Patrik Kolenovský, Vladimir Janisch · 2020
In this paper, an advanced variant of the efficient jDE100 variant is proposed. The introduced jDE100e employs an efficient rotationally-invariant Eigenvector crossover. Both algorithms are applied on 10 test problems for a single-objective optimisation CEC 2020 with four dimension levels. The results show that the proposed jDE100e is able to solve some of the problems successfully. A comparison of the proposed jDE100e with the original jDE100 illustrates the superiority of the newly designed algorithm as jDE100e performs significantly better in 16 out of 40 problems, and jDE100 is not able to outperform jDE100e significantly.