A Surrogate Multiobjective Evolutionary Strategy with Local Search and Pre-Selection

Autoři: Mgr. Roman Neruda, Ph.D.
Rok vydání: 2012
Typ publikace: Konferenční příspěvek (mezinárodní konf.)
Periodikum: GECCO Companion '12. Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference companion
Projekty: Analýza inteligentních distribuovaných výpočetních systémů
Bibliografická citace: PILÁT, M., NERUDA, Roman. A Surrogate Multiobjective Evolutionary Strategy with Local Search and Pre-Selection. In: SOULE, T., ed. GECCO Companion '12. Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference companion. New York: ACM, 2012, S. 633-634. ISBN 978-1-4503-1178-6
Dostupné z: doi:10.1145/2330784.2330895
DOI: 10.1145/2330784.2330895
ID: 380958

Abstrakt

In this paper we present an evolutionary strategy for multiobjective optimization. This evolution strategy is based on a surrogate memetic operator and a surrogate preselection model which provides several individuals in each generation. Thus, the optimization may be easily parallelized. The proposed algorithm is compared to some of existing evolutionary algorithms from the literature.