Yapay zeka tabanlı optimizasyon algoritmaları geliştirilmesi
Yükleniyor...
Dosyalar
Tarih
2017-07-11
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Selçuk Üniversitesi Fen Bilimleri Enstitüsü
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Yapay Zekâ (YZ), hiç kuşkusuz ki Bilgisayar Bilimleri'nin en önemli ilgi alanlarından birisidir. İnsan düşünce - davranış şekilleri ve doğa dinamiklerinin benzetimi neticesinde sunduğu etkin ve tutarlı çözümler nedeniyle hızlı bir şekilde gelişmiş ve modern yaşamın neredeyse her alanında yerini alarak gücünü kanıtlamıştır. Çok-disiplinli uygulanabilir olma özelliği, aynı zamanda farklı problem çeşitlerine kolaylıkla adapte edilebilmesiyle ilgilidir. Bu durum, YZ'nin kendi içerisinde alt-araştırma alanlarına ayrılmasına da sebep olmuştur. Optimizasyon, YZ'nin yoğun bir şekilde ilgili olduğu problem alanlarından birisidir. Genel olarak, eldeki kaynaklar bağlamında 'optimumu – en uygunu bulma' çabaları olarak tanımlayabileceğimiz optimizasyon, klasik tekniklerin yetersiz kalmaya başlaması sonrasında, çözümü YZ bünyesinde bulmuştur. Elde edilen başarılı sonuçlar, optimizasyon problemleri için kullanılan tekniklerin tasarlanmasına sebep olmuş ve sonuç olarak ilgili literatürde, YZ tabanlı optimizasyon tekniklerine yönelik bir alt-araştırma ortamı ortaya çıkmıştır. YZ tabanlı optimizasyon teknikleri tipik olarak çeşitli mantıksal ve matematiksel çözüm yaklaşımları çerçevesinde şekillenen ve çoğunlukla doğal dinamiklerden esinlenerek tasarlanan algoritmalar olarak bilinmektedir. Bu tür algoritmalara yönelik literatür özellikle son yıllarda büyük bir ivme kazanmış ve bilim insanlarının ilgi alanları içerisine girmiştir. Bu tez çalışmasının amacı, daha kolay kodlanabilen, alternatif YZ tabanlı optimizasyon algoritmaları geliştirmektir. Bu bağlamda, arkaplanı anlamak adına öncelikli olarak 'YZ ve optimizasyon' konusuna yönelik temel bilgiler sunulmuş ve literatürde ön plana çıkan bazı algoritmalardan bahsedilmiş, sonrasında ise, iki farklı sürekli optimizasyon algoritması geliştirilmiştir. Sırasıyla Girdap Optimizasyon Algoritması (GOA) ve Bilişsel Gelişim Optimizasyon Algoritması (BiGOA) adı verilen algoritmalar, Sürü Zekâsı adı altında incelenebilen çeşitli mekanizmaları ve basit matematiksel süreçleri içermektedir. Geliştirilen algoritmalar, çeşitli değerlendirme süreçlerinden geçirilmiş ve elde edilen bulgular, algoritmaların zeki optimizasyonda yeter düzeyde başarılı olduklarını göstermiştir.
It is clear that Artificial Intelligence (AI) is one of the most important interest areas of Computer Science. Because of its effective and efficient solutions provided as a result of efforts on simulating human thinking - behaviors and nature dynamics, it has improved rapidly and proved its power by taking place within almost all fields of the modern life. This feature of being applicable multidisciplinary is related to being also easily adaptable to different types of problems. This situation has also caused AI to be divided into sub-research fields. Optimization is one of the related problems to which the AI is related intensely. Optimization, which we can define generally as the efforts on 'finding the most appropriate – optimum one' in the sense of the sources in hand, has found its way in AI after classical techniques have become unsatisfactory on solutions. Obtained successful results have caused designing techniques used for optimization problems and as a result, a sub-research environment related to AI based optimization techniques has appeared. Typically, AI based optimization techniques are known as the algorithms formed in the context of various logical and mathematical solution approaches and mostly designed by inspiring from natural dynamics. The literature associated with such algorithms has gained momentum in especially recent years and taken part in scientists' interest areas. Objective of this thesis study is to develop alternative AI based optimization algorithms, which can be coded easier. In this sense, in order to understand the background, essential information regarding to the subject of 'AI and optimization' was given, and then some algorithms that come into prominence in the literature were explained primarily, and after that, two different continuous optimization algorithms were developed. Called as Vortex Optimization Algorithm (VOA), and Cognitive Development Optimization Algorithm (CoDOA), the related algorithms include various mechanism and some additional simple mathematical processes that can be examined in the context of Swarm Intelligence. The developed algorithms were employed in some evaluation processes and the findings showed that they are successful enough in intelligent optimization processes.
It is clear that Artificial Intelligence (AI) is one of the most important interest areas of Computer Science. Because of its effective and efficient solutions provided as a result of efforts on simulating human thinking - behaviors and nature dynamics, it has improved rapidly and proved its power by taking place within almost all fields of the modern life. This feature of being applicable multidisciplinary is related to being also easily adaptable to different types of problems. This situation has also caused AI to be divided into sub-research fields. Optimization is one of the related problems to which the AI is related intensely. Optimization, which we can define generally as the efforts on 'finding the most appropriate – optimum one' in the sense of the sources in hand, has found its way in AI after classical techniques have become unsatisfactory on solutions. Obtained successful results have caused designing techniques used for optimization problems and as a result, a sub-research environment related to AI based optimization techniques has appeared. Typically, AI based optimization techniques are known as the algorithms formed in the context of various logical and mathematical solution approaches and mostly designed by inspiring from natural dynamics. The literature associated with such algorithms has gained momentum in especially recent years and taken part in scientists' interest areas. Objective of this thesis study is to develop alternative AI based optimization algorithms, which can be coded easier. In this sense, in order to understand the background, essential information regarding to the subject of 'AI and optimization' was given, and then some algorithms that come into prominence in the literature were explained primarily, and after that, two different continuous optimization algorithms were developed. Called as Vortex Optimization Algorithm (VOA), and Cognitive Development Optimization Algorithm (CoDOA), the related algorithms include various mechanism and some additional simple mathematical processes that can be examined in the context of Swarm Intelligence. The developed algorithms were employed in some evaluation processes and the findings showed that they are successful enough in intelligent optimization processes.
Açıklama
Anahtar Kelimeler
Artificial intelligence, Yapay zeka, Heuristic algorithms, Sezgisel algoritmalar, Global optimization method, Global optimizasyon yöntemi, Evolutionary algorithms, Evrimsel algoritmalar
Kaynak
WoS Q Değeri
Scopus Q Değeri
Cilt
Sayı
Künye
Köse, U. (2017). Yapay zeka tabanlı optimizasyon algoritmaları geliştirilmesi. Selçuk Üniversitesi, Yayımlanmış doktora tezi, Konya.
