We propose a modification of the standard differential evolution (DE) algorithm in order to significantly make easier and more efficient standard DE implementations. Taking advantages from chaotic map approaches, recently proposed and successfully implemented for swarm intelligence-based algorithms, our DE improvement facilitates the search for the best population and then the optimal solution. More specifically, we work with a genetic memory that stores parents and grandparents of each individual (its kin) of the population generated by the DE algorithm. In this way, the new population is carried out not only on the basis of the best fitness of a certain individual, but also according to a good score of its kin. Additionally, we carried out a wide numerical campaign in order to assess the performances of our approach and validated the results with standard statistical techniques.
Formica, G., & Milicchio, F. (2019). Kinship-based differential evolution algorithm for unconstrained numerical optimization. NONLINEAR DYNAMICS [10.1007/s11071-019-05358-y].
|Titolo:||Kinship-based differential evolution algorithm for unconstrained numerical optimization|
|Data di pubblicazione:||2019|
|Citazione:||Formica, G., & Milicchio, F. (2019). Kinship-based differential evolution algorithm for unconstrained numerical optimization. NONLINEAR DYNAMICS [10.1007/s11071-019-05358-y].|
|Appare nelle tipologie:||1.1 Articolo in rivista|