A3-ARTIFICIAL ATOM ALGORITHM: A NEW META-HEURISTIC COMPUTATIONAL INTELLIGENCE ALGORITHM INSPIRED BY CHEMICAL PROCESSES
Küçük Resim Yok
Tarih
2018
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Ministry Communications & High Technologies Republic Azerbaijan
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
Most of the meta-heuristic algorithms are based on the natural processes. They were inspired by physical, biological, social, chemical, social-biological, biological-geography, music, and hybrid processes. In this paper, we propose a new meta-heuristic algorithm based on the chemical process of compound formation. In order to form compound, some electrons are removed from some elements, and some electrons are attached to some elements. The formed bond is called the ionic bond. In another case, at least two elements share at least one electron in common, and the constructed bond is called the covalent bond. Proposed algorithm based on compound formation process in this paper is a population based algorithm, and population consists of candidate solutions-subject to evolution for convergence to optimal/near optimal solutions. These solutions were called atoms, and the attributes of each solution were called electrons. The set of solutions constructs Atom Set where initially, atoms are constructed randomly. Then there are two operators in the proposed algorithm to make candidate solutions be better. So, this process is applied to solutions until desired solution is obtained.The formation of covalent bond in chemistry is mimicked as covalent bond where better attribute value is shared between at least two atoms. The ionic bond process is mimicked as ionic bond, and some attribute values of each atom reassigned randomly. The obtained algorithm was called as Artificial Atom Algorithm (A3). It is superior with respect to other natural-inspired metaheuristic algorithms, since it is applied once and global optimum or near-global optimum point was obtained. On contrary, other meta-heuristic algorithms are applied more than once, and the better obtained result is selected as a result of experiment. Another important point is that this algorithm does not have more algorithm parameters. There are just ionic rate (covalent rate) parameter (ionic rate +covalent rate = 1).
Açıklama
Anahtar Kelimeler
Meta-heuristic Methods, Evolutionary Computation, Genetic Algorithms, Nature-Inspired Methods
Kaynak
Applied and Computational Mathematics
WoS Q Değeri
Q1
Scopus Q Değeri
Q1
Cilt
17
Sayı
2
