AI Summary & Key Takeaways
AI InsightAs a result, it is shown that these algorithms gave better performances than the performances cited in the literature.
📌 3 Key Takeaways
Abstract
In this study, the parameters of a non-minimum phase second-order system with time delay, which is accepted as gray box, are found out using dual channel relay test and Cuckoo Search, Particle Swarm Optimization, Firefly Algorithms. For this purpose, gray box was implemented as two channel test systems to obtain signals belong to system inputs and outputs. After that, these signals are used to identify system parameter according to a performance criteria, which is integral absolute error between difference of real system and model outputs that was tried to be minimized by Cuckoo search, Particle Swarm Optimization and Firefly Algorithms. General system block diagram and its input and output are given at Figure A. Figure A. Two Channel System Block Diagram Purpose: Cuckoo Search, Particle Swarm Optimization and Firefly Algorithms are heuristic algorithms which are inspired from nature itself. Also these algorithm are capable of to solve benchmark problems. In this study a non-minimum phase second order system with time delay is tried to be identified by using Cuckoo Search, Particle Swarm Optimization and Firefly algorithms according to two channel relay test and performances of these algorithms are compared with each other for this problem. Theory and Methods: Firstly, a real system is implemented into the two channel relay test. This test allows the system to enter a graded oscillation in a limited range. Thus, different dynamics of the system can be stimulated. After that the model parameters are identified by the same test using Cuckoo Search, Particle Swarm Optimization and Firefly Algorithms. These algorithms were tried to minimize or maximize a performance criterion. In this problem, this criterion is integral absolute error between the model output and the system output that is tried to be minimized to identify the system parameters. Results: It is found that the obtained results using Cuckoo Search, Particle Swarm Optimization and Firefly Algorithms gave closer results to the real system than the results obtained with the Genetic Algorithm cited in the literature. Conclusion: In this study, a non-minimum phase second order system with time delay was tried to be identified using cuckoo search, particle swarm optimization and firefly algorithms with two channels relay test. Each algorithms run 10 times and their standard deviations and expected values were calculated. As a result, it is shown that these algorithms gave better performances than the performances cited in the literature.
Özet
Bu çalışmada, gri kutu olarak kabul edilen ikinci dereceden ölü zamanlı ve geri tepmeli bir sistemin parametreleri, çift röleli autotuning testi ve Cuckoo Search, Parçacık Sürü Optimizasyonu, FireFly Algoritmaları kullanılarak belirlenmiştir. Bunun için öncelikle gri kutu, çift röleli auto-tuning testine tabi tutularak, sistem giriş ve çıkışlarına ait sinyaller elde edilmiştir. Ardından bu giriş-çıkış sinyalleri arasındaki hata değeri, belirlenen amaç fonksiyonuna göre Cuckoo Search, Parçacık Sürü Optimizasyonu ve FireFly Algoritmaları kullanılarak minimize edilmekte ve sistem parametreleri belirlenmektedir. Amaç fonksiyonu olarak hatanın mutlak değerinin integrali(Integral Absolute Error) kriteri kullanılmıştır. Elde edilen sonuçlar analiz edilerek, literatürde var olan ve Genetik Algoritma ile yapılan çalışmalar ile kıyaslanmıştır. Cuckoo Search, Parçacık Sürü Optimizasyonu ve FireFly Algoritmaları kullanılarak bu çalışmada elde edilen sonuçların, Genetik Algoritma ile elde edilen sonuçlara göre gerçeğe daha yakın olduğu görülmüştür