A high-level and adaptive metaheuristic selection algorithm for solving high dimensional bound-constrained continuous optimization problems
Küçük Resim Yok
Tarih
2020
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Metaheuristic algorithms are used to find sufficiently good solutions for the optimization problems thatare not solvable in a polynomial time. Although metaheuristics offer a general problem-solving framework and canbe applied to various types of optimization problems, their performances depend heavily on the problem to be solved.Thus, hybrid metaheuristics are used to combine strong parts of different algorithms. In this study, a novel adaptivemetaheuristic selection algorithm is proposed for solving bound-constrained continuous optimization problems. Thedeveloped method hybridizes artificial bee colony, differential evolution, and particle swarm optimization at a high levelwhere each algorithm works independently from each other. As a main contribution to the literature, adaptive selectionat metaheuristic level among these three algorithms is achieved by using a rank-based credit assignment and UCB1multiarmed bandit selection. The effectiveness of the developed algorithm has been evaluated on CEC’17 standardbenchmark functions. The obtained numerical results indicate that the proposed algorithm outperforms the individualmetaheuristics on which it is built and is more effective especially in high dimensional problems. It is also shown thatthe proposed algorithm is highly comparable with the related algorithms in the literature. Lastly, a case study thatachieves adaptive selection of two good-performing algorithms (namely, covariance matrix adaptation evolution strategyand JADE) for the benchmark used in this study supports the effectiveness of the proposed method.
Açıklama
Anahtar Kelimeler
Kaynak
Turkish Journal of Electrical Engineering and Computer Sciences
WoS Q Değeri
Scopus Q Değeri
Cilt
28
Sayı
3
