Yazışmaların yapılacağı yazar: Bilal alataş. balatas
Yüklə 249,48 Kb. Pdf görüntüsü
|
- Bu səhifədəki naviqasiya:
- Tablo 7.
- Tablo 8.
- Sonuçlar ve Tartışma
- A new approach to swarm intelligence: Bird Swarm Algorithm Extended abstract
143 Sürü zekâsında yeni bir yaklaşım: Kuş sürüsü algoritması Tablo 7. Griewank fonksiyonunda PSO, YBSA, KSO, KAA ve KSA performanslarının karşılaştırılması P OPÜLASYON SAYISI ???????????????????????????????????? ????????????????????????????????????????????????????????????
???????????????????????????????????????????????? ????????????????????????????????????????????????????????????
100
2.4192 E -003 2.0008 E - 009 7.5301
E - 012 0.0405 0 50 7.7481 E -002 1.7812 E - 007 1.4704
E - 011 0.0642 0 30 3.4521 E -002 3.2304 E - 006 3.3501
E - 009 0.1913 0 10 1.0081 E -002 1.4102 E - 005 4.8516
E - 008 1.0382 0 2 2.1105 E -001 4.0529 E - 004 1.0063
E - 002 1.8185 0
[-5.12, 5.12] aralığında tanımlı 30 boyutlu Rastrigin fonksiyonunda farklı parçacık sayılı ve 1000 iterasyonlu dört optimizasyon algoritması arasındaki performans sonuçlarının değerlendirilmesi Tablo 8’de gösterilmiştir. Bu tabloya göre KSA; PSO, YBSA, KSO ve KAA’ya göre
çok modlu
Rastrigin fonksiyonunda çok daha iyi sonuçlar elde etmiştir.
KAA ve KSA performans karşılaştırılması P OPÜLASYON SAYISI ???????????????????????????????????? ????????????????????????????????????????????????????????????
???????????????????????????????????????????????? ????????????????????????????????????????????????????????????
???????????????????????????????????? KSA 100
1.08254 E - 004 7.3841
E -009
3.3557 E - 015 25.8705
0 50
2.3961 E - 002 1.0014
E -006
7.2471 E - 013 16.9198
0 30
8.73 4.5106
E -005
8.4519 E - 012 24.9272
0 10
10.26 9.0045
E -005
1.0041 E - 009 56.9561
0 2 14.73 4.78 2.12
1.22 0
Sonuçlar ve Tartışma Bu çalışmada sürü zekâsı optimizasyon algoritmalarının en güncellerinden biri olan kuş sürüsü optimizasyon algoritması ayrıntılı olarak incelenmiştir. Bu algoritmanın performansı, farklı boyutlardaki tek modlu çok modlu kalite testi fonksiyonları kullanılarak, yine güncel ve popüler sürü zekâsı algoritmalarından parçacık sürü optimizasyonu, yapay balık sürü optimizasyonu, kırlangıç sürü optimizasyonu ve karga arama algoritması ile karşılaştırılmıştır. Yapılan deneylerde, optimuma yakınsama eğilimi ve elde edilen sonuç değerleri, performans ölçütü olarak kullanılmıştır. İnceleme sonuçları karşılaştırmalı tablolar aracılığıyla sunulmuş ve yorumlanmıştır.
İnceleme için kalite testi fonksiyonları arasından en güncel olan dört fonksiyon seçilmiştir. Tek modlu fonksiyon olan Sphere fonksiyonunda KSA karşılaştırma için seçilen diğer dört sürü zekâsı algoritmalarından çok daha iyi sonuç elde etmiştir. Çok modlu fonksiyonlardan Griewank ve Rastrigin fonksiyonlarında KSA diğer dört algoritmadan çok daha iyi sonuç elde etmiştir. Sadece Rosenbrock fonksiyonunda istenilen başarı elde edilememiştir.
Genel olarak en güncel sürü zekâsı algoritmalarından biri olan KSA ile parametre ayarlaması yapılmadan hem tek modlu hem de çok modlu kalite testi fonksiyonlarında diğer sürü zekâsı algoritmalarından daha iyi sonuçlar elde edildiği için, algoritmanın ileride birçok kompleks problemde etkili olarak kullanılacağı beklenmektedir. KSA’nın
performansının arttırılması amacıyla; optimize edilmiş parametreler, dağıtık ve paralel versiyonlar, çok amaçlı çalışmalar, melez sistemler de geliştirilebilir. Kaynaklar Bunday B., (1984). Basic Optimization Methods, London: Edward Arnold Ltd. Kahaner D., Moler C., Nash S., (1989). Numerical Methods and Software. Englewood Cliffs: Prentice Hall. Himmelblau D.E., (11989). Optimization of Chemical Processes, NY: McGrawHill, Inc. Rao S., (1978). Optimization Theory and Applications, 2nd. Edition Halsted Inc. Altunbey F., Alataş B., (2015). Sosyal Ağ Analizi için Sosyal Tabanlı Yapay Zekâ Optimizasyon, IJPAS, 1(1), 33-52. Bonabeau E., Dorigo M. ve Theraulaz G., (1999). Swarm Intelligence: From Natural to Artificial System, New York: Oxford University Press. Kennedy J., Eberhart R., (1995). Particle Swarm Optimization, IEEE International Conference on Neural Networks, 1942-1948. Eberhart R.C., Kennedy J., (1995). A New Optimizer Using Particle Swarm Theory, Proceedings of the sixth international symposium on micro machine and human science. Vol. 1, 39- 43.
Karaboğa D., (2011). Yapay Zekâ Optimizasyon Algoritmaları, Nobel Yayın Dağıtım. Li X.L., Shao Z.J., Qian J.X., (2002). An Optimizing Method Based on Autonomous Animals: Fish- E. Varol, B. Alataş
Swarm Algorithm, Syst Eng Theory Pract 22(11):32–38. Zhang M., Shao C., Li F., Gan Y., Sun J., (2000) Evolving Neural Network Classifiers And Feature Subset Using Artificial fish Swarm, In: Proceedings of the 2006 IEEE International Conference on Mechatronics and Automation, June 25–28. Zhang M., Shao C., Li F., Gan Y., Sun J., (2007). Spread Spectrum Code Estimation by Artificial Fsh Swarm Algorithm, IEEE International Symposium on Intelligent Signal Processing (WISP). Mehdi N., Ghodrat S., Mehdi S. ve Adel N.T., (2012). Artificial Fish Swarm Algorithm: A Survey of the State-of-the-Art, Hybridization, Combinatorial and Indicative Applications, Springer Science. Bao L., Chu J., Liang Q., (2016). Artificial Fish Swarm Algorithm Based on Improved Topological Structure, Computer Science and Application, 6(3), 137-142. Duan Q., Mao M., Duan P., Hu B., (2016). An Improved Artificial Fish Swarm Algorithm Optimized by Particle Swarm Optimization Algorithm with Extended Memory, Kybernetes, 45(2): 210-222. Mehdi N., Ghodrat S. ve Mehdi S., (2012). Swallow Swarm Optimization Algorithm: A New Method to Optimization, Neural Computing and Applications. Meng X.B., Gao X.Z., Lu L., Liu Y., Zhang H., (2016). A New Bio-Inspired Optimisation Algorithm: Bird Swarm Algorithm, Journal of Experimental & Theoretical Artificial Intelligence, 28(4). Anderson T.R., (2006). Biology of the Ubiquitous House Sparrow: From Genes to Populations, Oxford University Press. Kennedy J., Eberhart R.C., Shi Y., (2001). Swarm Intelligence, Burlington: MA: Morgan Kaufmann. Krause J., Ruxton G.D., (2002). Living in Groups, Oxford University Press. Lima S.L., Dill L.M., (1990). Behavioral Decisions Made Under the Risk of Predation: A Review and Prospectus, Canadian Journal of Zoology, 68, 619–640. Bednekoff P.A., Lima S.L., (1998). Randomness, Chaos and Confusion in the Study of Antipredator Vigilance, Trends in Ecology and Evolution, 13, 284–287. Pulliam H.R., (1973). On the Advantages of Focking, Journal of Theoretical Biology, 38, 419–422. Ekman J., (1987). Exposure and Time Use in Willow Tit Focks: The cost of subordination, Animal Behavior, 35, 445–452. Beauchamp G., (1998). The Effect of Group Size on Mean Food Intake Rate in Birds, Biological Reviews, 73, 449–472. Roberts G., (2003). The Group-Size Effect in Non- feeding Animals, Behavioral Processes, 63, 127– 128. Beauchamp G., (2003). Group-size Effects on Vigilance: A search for mechanisms, Behavioral Processes, 63, 111–121. Barnard C.J., Sibly R.M., (1981). Producers and Scroungers: A General Model and Its Application to Captive Focks of House Sparrows, Animal Behavior, 29, 543–550. Giraldeau L.A., Caraco T., (2000). Social Foraging Theory, Princeton: Princeton University Press. Liker A., Barta Z., (2002). The Effects of Dominance on Social Foraging Tactic Use in House Sparrows, Behavior, 139, 1061–1076. Johnson C.A., Giraldeau L.A., Grant J.W., (2001). The Effect of Handling Time on Interference Among House Sparrows Foraging at Different Seed Densities, Behavior, 138, 597–614. Barta Z., Giraldeau L.A., (2000). Daily Patterns of Optimal Producer and Scrounger Use Under Predation Hazard: A State-Dependent Dynamic Game Analysis, American Naturalist, 155, 570– 582.
Sirot E., (2006). Social Information, Antipredatory Vigilance and Fight in Bird Focks, Animal Behavior, 72, 373–382. Askarzadeh A., (2016). A novel metaheuristic method for solving constrained engineering optimization problems: Crow search algorithm, Computers & Structures, 169, 1-12. Akyol S., Alatas B., (2016). Efficiency Evaluation of Crow Search Algorithm in Benchmark Functions for Optimization, ICENS 2016.
145 Sürü zekâsında yeni bir yaklaşım: Kuş sürüsü algoritması A new approach to swarm intelligence: Bird Swarm Algorithm Extended abstract Optimization known as also mathematical programming, is a collection of processes that selects the most appropriate values of decision variables according to a goal (evaluation) function. Many algorithms have been proposed for optimization problems. Most of these algorithms need mathematical models for model of system and objective function. Metaheuristics algorithms are the algorithms that utilize a simple approach as a solution technique of search and optimization problems and are recently getting strong and becoming more popular due to their advantages. They are population based techniques and begin to search the solution with multiple points. They have good reputation and wide acceptability as being powerful tools for many different fields such as management science, engineering, computer, etc. and new versions of these algorithms have been proposed. General purpose heuristic optimization algorithms are evaluated in eight different groups including biology-based, physics-based, swarm-based, social- based, music-based, chemical-based, sports based, and mathematics based.
continually searching the best and absence of the most efficient metaheuristic method for all types of problems, novel algorithms or new variants of current algorithms are being proposed. Swarm optimization algorithms are relatively newer subfield of computational intelligence and recently getting strong and becoming more popular. They have been developed by observing the movements of live swarms such as bird, fish, cat and bee. They are inspired from intelligent behaviors resulting from the local interactions of swarm agents between each other and environment. They are adaptable and general purposed solution methods that can be applied to the high-scale combinatorial and non- linear search and optimization problems in case of concurrent different decision variables, objective functions, and constraints and they do not depend on the solution space type, the number of decision variables, and the number of constraint functions. Furthermore, they do not require very well defined mathematical models that are hard to derive. Their computation power is also good and they do not require excessive computation time. Their transformations and adaptations are easy. Due to these advantages, these algorithms are densely being used in many different fields. In this work, bird swarm algorithm that is one of the most current swarm intelligence optimization algorithm was studied in detail. In this study, the performance of this new algorithm is tested within unimodal and multi modal benchmark functions with different dimensions. In these investigations, tendency of converging to optimum and the obtained result values are used as a measure of performance. Experimental results have been presented and interpreted through comparative tables. It is expected that this algorithm will be efficiently used in many different types of complex problems due to high performance of the algorithm in both unimodal and multi modal functions.
algorithms, Bird swarm algorithm, Benchmark functions. Yüklə 249,48 Kb. Dostları ilə paylaş:
|